PC GAMESS-related discussion club


Message d4c0f196bHW-5990-1130-00.htm, number 1, was posted on Fri May 26 '06 at 18:50:06
Welcome to the PC GAMESS discussion club!

Alex Granovsky
gran@classic.chem.msu.su


Welcome to the PC GAMESS discussion club!

Message c8d90b70ShW-5992-131-00.htm, number 2, was posted on Sun May 28 '06 at 02:11:45
inputs for transition metal calculations

silmar andrade
silmar.monte@uol.com.br


Dear all,

I would like to have as many as possible inputs concerning first transition series metal complexes.

Thanks in advance,

Silmar, PhD


Message d4763b0bbHW-5994-1102+00.htm, number 3, was posted on Tue May 30 '06 at 18:22:07
in reply to c8d90b70ShW-5992-131-00.htm

Re: inputs for transition metal calculations

Alex Granovsky
gran@classic.chem.msu.su


Dear Silmar,

do you have any specific problems with your calculations?
Please describe them in more details, if so. Or, do you need just
some general advices?

P.S. Note that this forum was established only several days ago
so it is wise to send questions to gamess list as well. However,
it is possible to upload input files, etc... to this forum, while
gamess list refuses attachments.


On Sun May 28 '06, silmar andrade wrote
---------------------------------------
>Dear all,

>I would like to have as many as possible inputs concerning first transition series metal complexes.

>Thanks in advance,

>Silmar, PhD


Message c1e879b1VYm-6000-872-00.htm, number 4, was posted on Mon Jun 5 '06 at 14:31:54
Parallel MP2 gradient

Alex Kargovsky
kargovsky@yumr.phys.msu.su


Dear PC GAMESS users!

Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
Then I try to do these calculation type, I get this:

PARALLEL MP2 GRADIENT IS VERBOTEN

Here a my input:
$CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
$SYSTEM MWORDS=105 timlim=6000 $END
$scf dirscf=.t. fdiff=.f. $end
$STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
$force nvib=2 $end
$data
...

Thanks in advance


Message c1e879b1VYm-6000-872+00.htm, number 4, was edited on Mon Jun 5 '06 at 14:32:54
and replaces message c1e879b1VYm-6000-872-00.htm

Parallel MP2 gradient

Alex Kargovsky
kargovsky@yumr.phys.msu.su


Dear PC GAMESS users!

Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
Then I try to do these calculation type, I get this:

PARALLEL MP2 GRADIENT IS VERBOTEN

Here a my input:
$CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
$SYSTEM MWORDS=105 timlim=6000 $END
$scf dirscf=.t. fdiff=.f. $end
$STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
$force nvib=2 $end
$data
...

Thanks in advance

[ This message was edited on Mon Jun 5 '06 by the author ]


Message d4c0f196bHW-6000-1151+00.htm, number 5, was posted on Mon Jun 5 '06 at 19:10:49
in reply to c1e879b1VYm-6000-872+00.htm

Re: Parallel MP2 gradient

Alex Granovsky
gran@classic.chem.msu.su


Dear Alex,

at present, there is no parallel MP2 gradient code in the PC GAMESS.
Note, parallel MP2 properties are supported. We are working on
brand new, low memory demand, efficient highly scalable
parallel MP2 gradient code. Most likely, it will become available
in the beta versions of the forthcoming PC GAMESS v. 7.1 and
will be included into the final release of the PC GAMESS v. 7.1

On Mon Jun 5 '06, Alex Kargovsky wrote
--------------------------------------
>Dear PC GAMESS users!

>Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
>Then I try to do these calculation type, I get this:

>PARALLEL MP2 GRADIENT IS VERBOTEN

>Here a my input:
> $CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
> $SYSTEM MWORDS=105 timlim=6000 $END
> $scf dirscf=.t. fdiff=.f. $end
> $STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
> $force nvib=2 $end
> $data
>...

>Thanks in advance


Message d4c0f196bHW-6004-1212+00.htm, number 6, was posted on Fri Jun 9 '06 at 20:11:59
in reply to d4c0f196bHW-5990-1130-00.htm

Welcome to the PC GAMESS discussion club!

Alex Granovsky
gran@classic.chem.msu.su


We have finished configuration of the PC GAMESS Forum pages. All options (including email notification) are working now.


Message c22c517caGH-6056-1275-00.htm, number 7, was posted on Mon Jul 31 '06 at 21:16:09
PC GAMESS ignores FVALUE?

Roman Zubatyuk
zubatyuk@gmail.com


It seems that the latest versions of the program ignores FVALUE keyword in $zmat. I'm trying to freeze internal coordinate to 6.0 using IFZMAT and FVALUE according to common GAMESS (US)/PC GAMESS documentation. In the input geometry this internal coordinate has value of 6.07. The PC GAMESS actually freezes the coordinate, but does not change its value to specified by FVALUE keyword. Is there any new keywords that have replaced the FVALUE?

Here is my $ZMAT group:
INPUT CARD> $zmat dlc=.t. auto=.t.
INPUT CARD> ifzmat(1)=1,18,29 fvalue(1)=6.0 nonvdw(1)=18,29 $end

Roman.


Message d4763b0bbHW-6057-1205+00.htm, number 8, was posted on Tue Aug 1 '06 at 20:05:34
in reply to c22c517caGH-6056-1275-00.htm

Re: PC GAMESS ignores FVALUE?

Alex Granovsky
gran@classic.chem.msu.su


On Mon Jul 31 '06, Roman Zubatyuk wrote
---------------------------------------
>It seems that the latest versions of the program ignores FVALUE keyword in $zmat. I'm trying to freeze internal coordinate to 6.0 using IFZMAT and FVALUE according to common GAMESS (US)/PC GAMESS documentation. In the input geometry this internal coordinate has value of 6.07. The PC GAMESS actually freezes the coordinate, but does not change its value to specified by FVALUE keyword. Is there any new keywords that have replaced the FVALUE?

>Here is my $ZMAT group:
> INPUT CARD> $zmat dlc=.t. auto=.t.
> INPUT CARD> ifzmat(1)=1,18,29 fvalue(1)=6.0 nonvdw(1)=18,29 $end

>Roman.
>

The following is a very brief description of several new keywords of the $ZMAT group.

AUTOFV (default .true.) means to generate fvalue array automatically. Any user input of fvalue is thus ignored. This option is especially useful for relaxed PES scans in DLC.

DLCTOL (default 1.0d-5) is the threshold used at several stages during DLC generation, as well as to impose geometrical constrains. Normally you do not need to alter it, however, in some cases setting it to, say, 1d-7 can help generating required number of linearly independent DLC coordinates.

IFRZAT is the array of atom numbers you want to freeze during geometry optimization using DLC. More precisely, the distances between all chosen atoms will be frozen, however, they will be allowed to rotate and translate as the single, united group (rigid body).

Hope this helps
Alex Granovsky


Message d4763b0bbHW-6071-1043-00.htm, number 9, was posted on Tue Aug 15 '06 at 17:22:52
New PC GAMESS benchmarks data

Alex Granovsky
gran@classic.chem.msu.su


We have added several new pages, including

Everybody are welcome to discuss results.

Message c22c517caGH-6071-1278+00.htm, number 10, was posted on Tue Aug 15 '06 at 21:18:10
in reply to d4763b0bbHW-6071-1043-00.htm

Re: New PC GAMESS benchmarks data

Roman Zubatyuk
zubatyuk@gmail.com


PC GAMESS v. 7.0 benchmarks on different Intel Pentium 4 family (Netburst) processors

The wall clock time is only given for large semi-empirical job. But I think that speedup on dual-core Pentium with respect to single-CPU computers is very interesting point. As for now (I guess) dual-core machines have optimal price/performance ratio. From the other hand, simultaneous access of two processes to the shared memory and disk can be a bottleneck on a some kinds of jobs. Please, can you give also wall clock times.

Best regards,
Roman.

On Tue Aug 15 '06, Alex Granovsky wrote
---------------------------------------
>We have added several new pages, including
>


>Everybody are welcome to discuss results.

Message d4c0ee0abHW-6071-1366+00.htm, number 11, was posted on Tue Aug 15 '06 at 22:46:56
in reply to c22c517caGH-6071-1278+00.htm

Re^2: New PC GAMESS benchmarks data

Alex Granovsky
gran@classic.chem.msu.su


The Wall clock times can be found in the corresponding output files. Typically, they are close to CPU times. The only exception is Test #2, which is partially limited by disk I/O on systems with relatively small amount of installed RAM. This is the only reason why both CPU and Wall clock times are provided in this case.

Best regards,
Alex Granovsky

On Tue Aug 15 '06, Roman Zubatyuk wrote
---------------------------------------
>PC GAMESS v. 7.0 benchmarks on different Intel Pentium 4 family (Netburst) processors

> The wall clock time is only given for large semi-empirical job. But I think that speedup on dual-core Pentium with respect to single-CPU computers is very interesting point. As for now (I guess) dual-core machines have optimal price/performance ratio. From the other hand, simultaneous access of two processes to the shared memory and disk can be a bottleneck on a some kinds of jobs. Please, can you give also wall clock times.

>Best regards,
> Roman.

>On Tue Aug 15 '06, Alex Granovsky wrote
>---------------------------------------
>>We have added several new pages, including
>>


>>Everybody are welcome to discuss results.

Message c22c517caGH-6074-1315-00.htm, number 12, was posted on Fri Aug 18 '06 at 21:55:42
External basis set file

Roman Zubatyuk
zubatyuk@gmail.com


Dear PC GAMESS developers,

At the readme.commandline file at the PC GAMESS web site (but not at the PC GAMESS distribution archive) the -b option mentioned to use external basis set library file. I was failed to find more information about this in GAMESS archives and PC GAMESS documentation. Is use of external basis set library the same as for GAMESS: exfil=.True. and gbasis=xxxxxx in $basis ? And what about format of the basis set library file? If the format is the same as for GAMESS, I can provide complete library file with Dunning's (Aug)-cc-pVNZ and possibly with Ahlrich's def2 basis sets.

By the way, at the announce message about PC GAMESS 7.0.1 you have mentioned that this new version has some new features. Is this capability to use external basis set file the new feature? Is this the only new feature?

Best regards,
Roman.


Message d4c0ee0abHW-6074-1407+00.htm, number 13, was posted on Fri Aug 18 '06 at 23:27:57
in reply to c22c517caGH-6074-1315-00.htm

Re: External basis set file

Alex Granovsky
gran@classic.chem.msu.su


Dear Roman!

> At the readme.commandline file at the PC GAMESS web site
>(but not at the PC GAMESS distribution archive) the -b option
> mentioned to use external basis set library file. I was failed to
> find more information about this in GAMESS archives and PC GAMESS
> documentation. Is use of external basis set library the same as
> for GAMESS: exfil=.True. and gbasis=xxxxxx in $basis ?


This option is implemented and works exactly as in the GAMESS (US). This means you need to add

$basis extfil=.t. gbasis=name $end

to your input to activate it.

This option was added starting from one of the multiple PC GAMESS v. 6.5 Release Candidates about one year ago. The format of basis set files is exactly the same as in GAMESS (US) as well, e.g.:

!Supported elements
!cc-pVTZ:
!H  He Li Be B  C  N  O  F  Ne Na Mg Al Si P  S  Cl Ar Ca Ga Ge As Se 
!Br Kr                                                                


H cc-pVTZ
 S   3
  1        33.87000000         0.6068000000E-02
  2        5.095000000         0.4530800000E-01
  3        1.159000000         0.2028220000    
 S   1
  1       0.3258000000          1.000000000    
 S   1
  1       0.1027000000          1.000000000    
 P   1
  1        1.407000000          1.000000000    
 P   1
  1       0.3880000000          1.000000000    
 D   1
  1        1.057000000          1.000000000    




By default, PC GAMESS looks for the file called basis.lib in its working directory. If the -b < path to basisfile > option is given, then it looks for file called "path to basisfile" first. If it does not exist, it then tries to find file called basis.lib in the directory "path to basisfile".

> for GAMESS, I can provide complete library file with Dunning's
> (Aug)-cc-pVNZ and possibly with Ahlrich's def2 basis sets.

Thanks so much for suggestion. Please send them directly to me and we'll put them on the PC GAMESS homepage.


> By the way, at the announce message about PC GAMESS 7.0.1 you have
> mentioned that this new version has some new features. Is this
> capability to use external basis set file the new feature? Is this
> the only new feature?

Note, the main emphasis in this release was on Core2 optimization.

Extfil is not the new feature. Briefly, new features are the optional use of fastints-based CAS-style integral transformation for GUGA transition dipole and spin-orbit calculations, printout of CIS & TD transition dipoles between states and states dipole moments, the improved stability of CIS CG AO CPHF solver and several options to fine control it. Several minor bugs (inherited from GAMESS (US)) were fixed as well.

Best regards,
Alex Granovsky


Message d4c0ee0abHW-6075-735+00.htm, number 13, was edited on Sat Aug 19 '06 at 12:15:24
and replaces message d4c0ee0abHW-6074-1407+00.htm

Re: External basis set file

Alex Granovsky
gran@classic.chem.msu.su


Dear Roman!

> At the readme.commandline file at the PC GAMESS web site
>(but not at the PC GAMESS distribution archive) the -b option
> mentioned to use external basis set library file. I was failed to
> find more information about this in GAMESS archives and PC GAMESS
> documentation. Is use of external basis set library the same as
> for GAMESS: exfil=.True. and gbasis=xxxxxx in $basis ?


This option is implemented and works exactly as in the GAMESS (US). This means you need to add

$basis extfil=.t. gbasis=name $end

to your input to activate it.

This option was added starting from one of the multiple PC GAMESS v. 6.5 Release Candidates about one year ago. The format of basis set files is exactly the same as in GAMESS (US) as well, e.g.:

!The following is just a comment
!Supported elements
!cc-pVTZ:
!H  He Li Be B  C  N  O  F  Ne Na Mg Al Si P  S  Cl Ar Ca Ga Ge As Se 
!Br Kr                                                                


H cc-pVTZ
 S   3
  1        33.87000000         0.6068000000E-02
  2        5.095000000         0.4530800000E-01
  3        1.159000000         0.2028220000    
 S   1
  1       0.3258000000          1.000000000    
 S   1
  1       0.1027000000          1.000000000    
 P   1
  1        1.407000000          1.000000000    
 P   1
  1       0.3880000000          1.000000000    
 D   1
  1        1.057000000          1.000000000    




By default, PC GAMESS looks for the file called basis.lib in the directory where input file resides. If the -b < path to basisfile > option is given, then it looks for file called "path to basisfile" first. If it does not exist, it then tries to find file called basis.lib in the directory "path to basisfile".

> for GAMESS, I can provide complete library file with Dunning's
> (Aug)-cc-pVNZ and possibly with Ahlrich's def2 basis sets.

Thanks so much for suggestion. Please send them directly to me and we'll put them on the PC GAMESS homepage.


> By the way, at the announce message about PC GAMESS 7.0.1 you have
> mentioned that this new version has some new features. Is this
> capability to use external basis set file the new feature? Is this
> the only new feature?

Note, the main emphasis in this release was on Core2 optimization.

Extfil is not the new feature. Briefly, new features are the optional use of fastints-based CAS-style direct integral transformation for GUGA transition dipole and spin-orbit calculations, printout of CIS & TD transition dipoles between states and states dipole moments, the improved stability of CIS CG AO CPHF solver and several options allowing to fine control it. Several minor bugs (inherited from GAMESS (US)) were fixed as well.

Best regards,
Alex Granovsky

[ This message was edited on Sat Aug 19 '06 by the author ]


Message c1891821bQn-6085-1075+00.htm, number 14, was posted on Tue Aug 29 '06 at 17:54:48
in reply to d4763b0bbHW-6071-1043-00.htm

Re: New PC GAMESS benchmarks data

Pedro Silva
pedros@ufp.pt


In some of the tests you reported, you detected bugs in Wingamess. How do we know for sure that the PcGamess result is the right one?

Pedro S


Message c1891821bQn-6085-1090-00.htm, number 15, was posted on Tue Aug 29 '06 at 18:09:43
PCM model

Pedro Silva
pedros@ufp.pt


Current Gamess-US documentation states that prior to Oct. 2000 D-PCM was the only PCM model in the program. I assume that since PcGamess branched from Gamess-US before that time, D-PCM is also the only PCM model there, too. According to

Barone V, Improta R, Rega N
Computation of protein pK's values by an integrated density functional theory/Polarizable Continuum Model approach
THEORETICAL CHEMISTRY ACCOUNTS 111 (2-6): 237-245 MAR 2004


C-PCM is far superior to D-PCM (unless compensation schemes for escaped charges are incorporated in D-PCM). PcGamess output tells me that escaped charge effects are taken into account, but no details are given. What is the compensation scheme used?



Thank you very much!
Pedro S.


Message d4763b02bHW-6085-1152+00.htm, number 16, was posted on Tue Aug 29 '06 at 19:12:15
in reply to c1891821bQn-6085-1075+00.htm

Re^2: New PC GAMESS benchmarks data

Alex Granovsky
gran@classic.chem.msu.su


Because results were compared with those of other QC packages.


On Tue Aug 29 '06, Pedro Silva wrote
------------------------------------
>In some of the tests you reported, you detected bugs in Wingamess. How do we know for sure that the PcGamess result is the right one?

>Pedro S


Message d4763b02bHW-6085-1169+00.htm, number 17, was posted on Tue Aug 29 '06 at 19:28:59
in reply to c1891821bQn-6085-1090-00.htm

Re: PCM model

Alex Granovsky
gran@classic.chem.msu.su


1. B.Mennucci, J.Tomasi J.Chem.Phys. 106, 5151-5158(1997)

2. Another paper in JCC


On Tue Aug 29 '06, Pedro Silva wrote
------------------------------------
>Current Gamess-US documentation states that prior to Oct. 2000 D-PCM was the only PCM model in the program. I assume that since PcGamess branched from Gamess-US before that time, D-PCM is also the only PCM model there, too. According to

>Barone V, Improta R, Rega N
>Computation of protein pK's values by an integrated density functional theory/Polarizable Continuum Model approach
>THEORETICAL CHEMISTRY ACCOUNTS 111 (2-6): 237-245 MAR 2004
>
>
>C-PCM is far superior to D-PCM (unless compensation schemes for escaped charges are incorporated in D-PCM). PcGamess output tells me that escaped charge effects are taken into account, but no details are given. What is the compensation scheme used?
>
>
>
>Thank you very much!
>Pedro S.


Message d4763b02bHW-6085-1192+00.htm, number 16, was edited on Tue Aug 29 '06 at 19:52:26
and replaces message d4763b02bHW-6085-1152+00.htm

Re^2: New PC GAMESS benchmarks data

Alex Granovsky
gran@classic.chem.msu.su


Because results were compared with those of other QC packages by Dr. Popov who provided us by these data.


On Tue Aug 29 '06, Pedro Silva wrote
------------------------------------
>In some of the tests you reported, you detected bugs in Wingamess. How do we know for sure that the PcGamess result is the right one?

>Pedro S

[ This message was edited on Tue Aug 29 '06 by the author ]


Message 589d47b6bQn-6085-1379+00.htm, number 18, was posted on Tue Aug 29 '06 at 22:59:31
in reply to d4763b02bHW-6085-1192+00.htm

Re^3: New PC GAMESS benchmarks data

Pedro Silva
pedros@ufp.pt


Thanks! I hope you did not consider my question rude: I just get asked that question a lot by colleagues with more $$$ who use Gaussian....


Pedro

On Tue Aug 29 '06, Alex Granovsky wrote
---------------------------------------
>Because results were compared with those of other QC packages by Dr. Popov who provided us by these data.
>
>


Message c17d4ed5eOC-6112-1090-00.htm, number 19, was posted on Mon Sep 25 '06 at 18:10:37
Surface and BSSE

Solntsev Pasha
solntsev@univ.kiev.ua


Is it possible to perform surface scan and BSSE correction and were can i read more about it? What is the clear-cut distinction between "rsurface" and "surface" key words?

Message d4763b03bHW-6113-1096+00.htm, number 20, was posted on Tue Sep 26 '06 at 18:16:37
in reply to c17d4ed5eOC-6112-1090-00.htm

Re: Surface and BSSE

Alex Granovsky
gran@classic.chem.msu.su


If you mean scan with BSSE correction included into it, this is not programmed at present. However, you can separately perform scans and separately calculate BSSE correction (either manually using CP method, or using RUNTYP=MOROKUMA).

The difference between "rsurface" and "surface" is that the relaxed scans allow relaxation of the molecular geometry (over those degrees of freedom which are not involved into scan) while scanning. Surface is the much more simple scanning over one or two degrees of freedom using frozen (initial) values for the rest of variables.

Best regards,
Alex Granovsky

On Mon Sep 25 '06, Solntsev Pasha wrote
---------------------------------------
>Is it possible to perform surface scan and BSSE correction and were can i read more about it? What is the clear-cut distinction between "rsurface" and "surface" key words?


Message db6eef3fcRR-6125-843-00.htm, number 21, was posted on Sun Oct 8 '06 at 14:03:20
In the case of "too big system" for shells reordering

Y. Matsubara
ymatsuba@chem.titech.ac.jp


Dear PC GAMESS users,


 In the energy calculation of a large molecular having C2h symmetry,
PC GAMESS 7.0.1 for x86 PC output error messages as follows when total number of basis set shells are 1572, for example.

     RUN TITLE
     ---------
 PC GAMESS

 THE POINT GROUP OF THE MOLECULE IS CNH
 THE ORDER OF THE PRINCIPAL AXIS IS     2

 FATAL: THE SYSTEM IS TOO BIG FOR SHELLS REORDERING.

In the other hand, GAMESS (US) version February 22, 2006 for Microsoft Windows begins to calculate normally.
Does anyone have a good idea for this problem?


Sincerely,
Yasuo Matsubara

--
INPUT is as follows:
 $CONTRL SCFTYP=RHF RUNTYP=energy COORD=UNIQUE MAXIT=200 NZVAR=0
         ICHARG=0 exetyp=run $END
 $SYSTEM TIMLIM=6000 MEMORY=50000000 $END
 $STATPT NSTEP=500 OPTTOL=0.0001 $END
 $BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END
 $SCF DIRSCF=.F. DAMP=.T. DIIS=.T. $END
 $GUESS GUESS=HUCKEL  $END
 $DATA
PC GAMESS
Cnh 2

    C    6    8.2820    0.0000    0.0000
...
(136 atoms)
...
 $END

WinGamess version 22nd February 2006 output as follows:
...

   1572   S    1119             0.1612778    1.000000000000

 TOTAL NUMBER OF BASIS SET SHELLS             = 1572
 NUMBER OF CARTESIAN GAUSSIAN BASIS FUNCTIONS = 4710
 NUMBER OF ELECTRONS                          = 2047
 CHARGE OF MOLECULE                           =    0
 SPIN MULTIPLICITY                            =    1
 NUMBER OF OCCUPIED ORBITALS (ALPHA)          = 1023
 NUMBER OF OCCUPIED ORBITALS (BETA )          = 1023
 TOTAL NUMBER OF ATOMS                        =  498
 THE NUCLEAR REPULSION ENERGY IS   113007.2708103098

...

Message d4763b0bbHW-6125-933+00.htm, number 22, was posted on Sun Oct 8 '06 at 15:32:45
in reply to db6eef3fcRR-6125-843-00.htm

Re: In the case of "too big system" for shells reordering

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this message means that your system is too large for shells reordering using by PC GAMESS by default.
Add $contrl reordr=.f. $end to your input to disable reordering. This will slightly slow-down direct SCF but will allow you to proceed.

Best regards,
Alex Granovsky



On Sun Oct 8 '06, Y. Matsubara wrote
------------------------------------
>

Dear PC GAMESS users,


 In the energy calculation of a large molecular having C2h symmetry,
PC GAMESS 7.0.1 for x86 PC output error messages as follows when total number of basis set shells are 1572, for example.

     RUN TITLE
     ---------
 PC GAMESS

 THE POINT GROUP OF THE MOLECULE IS CNH
 THE ORDER OF THE PRINCIPAL AXIS IS     2

 FATAL: THE SYSTEM IS TOO BIG FOR SHELLS REORDERING.

In the other hand, GAMESS (US) version February 22, 2006 for Microsoft Windows begins to calculate normally.
Does anyone have a good idea for this problem?


Sincerely,
Yasuo Matsubara

--
INPUT is as follows:
 $CONTRL SCFTYP=RHF RUNTYP=energy COORD=UNIQUE MAXIT=200 NZVAR=0
         ICHARG=0 exetyp=run $END
 $SYSTEM TIMLIM=6000 MEMORY=50000000 $END
 $STATPT NSTEP=500 OPTTOL=0.0001 $END
 $BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END
 $SCF DIRSCF=.F. DAMP=.T. DIIS=.T. $END
 $GUESS GUESS=HUCKEL  $END
 $DATA
PC GAMESS
Cnh 2

    C    6    8.2820    0.0000    0.0000
...
(136 atoms)
...
 $END

WinGamess version 22nd February 2006 output as follows:
...

   1572   S    1119             0.1612778    1.000000000000

 TOTAL NUMBER OF BASIS SET SHELLS             = 1572
 NUMBER OF CARTESIAN GAUSSIAN BASIS FUNCTIONS = 4710
 NUMBER OF ELECTRONS                          = 2047
 CHARGE OF MOLECULE                           =    0
 SPIN MULTIPLICITY                            =    1
 NUMBER OF OCCUPIED ORBITALS (ALPHA)          = 1023
 NUMBER OF OCCUPIED ORBITALS (BETA )          = 1023
 TOTAL NUMBER OF ATOMS                        =  498
 THE NUCLEAR REPULSION ENERGY IS   113007.2708103098

...

Message 83701356cRR-6125-1083+00.htm, number 23, was posted on Sun Oct 8 '06 at 18:02:55
in reply to d4763b0bbHW-6125-933+00.htm

Re^2: In the case of "too big system" for shells reordering

Y. Matsubara
ymatsuba@chem.titech.ac.jp


Thank you for quick comment. As I turned off reorder, this problem was solved.
But, PC GAMESS stopped on the next problem as follows:
 ...
 ***   D 1127       0.800000    1.113825 (  1.000000) 

 H         

 ***   S 1128      18.731137    0.214935 (  0.033495) 
 ***   S 1129       2.825394    0.364571 (  0.234727) 
 ***   S 1130       0.640122    0.415051 (  0.813757) 

 ***   S 1131       0.161278    0.181381 (  1.000000) 
 *** ERROR, NO MORE THAN  4095 TOTAL BASIS FUNCTIONS

I didn't found the option for the number of basis functions in the PC GAMESS documentation.
What should I do? Should I turn on QFMM?

Sincerely,
Yasuo MATSUBARA


On Sun Oct 8 '06, Alex Granovsky wrote
--------------------------------------
>Hi,
>
>this message means that your system is too large for shells >reordering using by PC GAMESS by default.
>Add $contrl reordr=.f. $end to your input to disable reordering. >This will slightly slow-down direct SCF but will allow you to >proceed.
>
>Best regards,
>Alex Granovsky


Message db6eef3fcRR-6125-1227+00.htm, number 24, was posted on Sun Oct 8 '06 at 20:26:53
in reply to d4763b0bbHW-6125-933+00.htm

Re^2: In the case of "too big system" for shells reordering

Y. Matsubara
ymatsuba@chem.titech.ac.jp


Thank you for quick comment. As I turned off reorder, this problem was solved. But, PC GAMESS stopped on the next problem as follows:
 ...
 ***   D 1127       0.800000    1.113825 (  1.000000) 

 H         

 ***   S 1128      18.731137    0.214935 (  0.033495) 
 ***   S 1129       2.825394    0.364571 (  0.234727) 
 ***   S 1130       0.640122    0.415051 (  0.813757) 

 ***   S 1131       0.161278    0.181381 (  1.000000) 
 *** ERROR, NO MORE THAN  4095 TOTAL BASIS FUNCTIONS

I didn't found the option for the number of basis functions in the PC GAMESS documentation. Should I turn on QFMM?

Sincerely,
Yasuo MATSUBARA


On Sun Oct 8 '06, Alex Granovsky wrote
--------------------------------------
>Hi,
>
>this message means that your system is too large for shells
>reordering using by PC GAMESS by default.
>Add $contrl reordr=.f. $end to your input to disable reordering.
>This will slightly slow-down direct SCF but will allow you to >proceed.
>
>Best regards,
>Alex Granovsky


Message d4763b03bHW-6125-1238+00.htm, number 25, was posted on Sun Oct 8 '06 at 20:38:31
in reply to db6eef3fcRR-6125-1227+00.htm

Re^3: In the case of "too big system" for shells reordering

Alex Granovsky
gran@classic.chem.msu.su


The limit of 4095 basis functions is presently hardcoded in some parts of the PC GAMESS. Thus, you cannot use more than 4095 basis functions without code changes and recompiling. Moreover, fast MP2 energy code will not work with more than 4095 basis functions as currently implemented.

If you really need to use more than 4096 basis functions, we can consider creation of customized version of PC GAMESS for you.

On Sun Oct 8 '06, Y. Matsubara wrote
------------------------------------
>Thank you for quick comment. As I turned off reorder, this problem was solved. But, PC GAMESS stopped on the next problem as follows:
>

 ...
 ***   D 1127       0.800000    1.113825 (  1.000000) 

 H         

 ***   S 1128      18.731137    0.214935 (  0.033495) 
 ***   S 1129       2.825394    0.364571 (  0.234727) 
 ***   S 1130       0.640122    0.415051 (  0.813757) 

 ***   S 1131       0.161278    0.181381 (  1.000000) 
 *** ERROR, NO MORE THAN  4095 TOTAL BASIS FUNCTIONS

>I didn't found the option for the number of basis functions in the PC GAMESS documentation. Should I turn on QFMM?

>Sincerely,
>Yasuo MATSUBARA


Message db6eef3fcRR-6126-518+00.htm, number 26, was posted on Mon Oct 9 '06 at 08:38:17
in reply to d4763b03bHW-6125-1238+00.htm

Re^4: In the case of "too big system" for shells reordering

Y. Matsubara
ymatsuba@chem.titech.ac.jp


I understood the limitation of the number of basis functions.

Since I have some XRD data of the cyclic metal complexes such as this large molecular, which has physical properties worth calculating, I try calculating about slightly "light" molecular than this molecular.


Sincerely,
Yasuo MATSUBARA


On Sun Oct 8 '06, Alex Granovsky wrote
--------------------------------------
The limit of 4095 basis functions is presently hardcoded in some parts of the PC GAMESS. Thus, you cannot use more than 4095 basis functions without code changes and recompiling. Moreover, fast MP2 energy code will not work with more than 4095 basis functions as currently implemented.

If you really need to use more than 4096 basis functions, we can consider creation of customized version of PC GAMESS for you.


Message db6eef89cRR-6126-767+00.htm, number 27, was posted on Mon Oct 9 '06 at 12:47:09
in reply to d4763b03bHW-6125-1238+00.htm

Re^4: In the case of "too big system" for shells reordering

Y. Matsubara
ymatsuba@chem.titech.ac.jp


I understood the limitation of the number of basis functions in PC GAMESS 7.0.1.

Since I have some XRD data of the cyclic metal complexes such as this large molecular, which has physical properties worth calculating, I try calculating about slightly "lighter" molecular than this molecular first.


Sincerely,
Yasuo MATSUBARA


On Sun Oct 8 '06, Alex Granovsky wrote
--------------------------------------
The limit of 4095 basis functions is presently hardcoded in some parts of the PC GAMESS. Thus, you cannot use more than 4095 basis functions without code changes and recompiling. Moreover, fast MP2 energy code will not work with more than 4095 basis functions as currently implemented.

If you really need to use more than 4096 basis functions, we can consider creation of customized version of PC GAMESS for you.


Message c8027773Sma-6129-150-00.htm, number 28, was posted on Thu Oct 12 '06 at 02:30:26
PES with rsurface

Francisco Burgos
franciscoburgos@uach.cl


Hi PCGamess users.
I am a little child using gamess and pcgamess, my question is about relaxes potential energy surfaces (PES) that is implemented in pcgamess whit rsurface. When I tried to do an scan of a torsion angle, starting on zero degrees, begin very well, but in the second step pcgamess take as input the initial coordenates, giveb ai $DATA, but with the necesary changed angle. This is not problem for the first step, but, when the running is advanced the initial coordenates are very poor and the calculations fails, at scf level or statpt level.
Are there a way to going passing the contrained optimized structure from the step n-1 to the n-th step and so on?

The output file is very hugh!!!! (63Mb) and at the end of this run fails with a this legende:


MAXIMUM GRADIENT = 45.3078979 RMS GRADIENT = 7.4281827

0OPTIMIZATION ABORTED.

-- GRADIENT OUT OF RANGE

-- MAXIMUM ALLOWED FORCE (FMAXT) = 10.000





EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 21:46:15 LT 11-OCT-2006

520354 WORDS OF DYNAMIC MEMORY USED



CPU TIME: STEP = 0.02 , TOTAL = 431.0 SECONDS ( 7.2 MIN)

WALL CLOCK TIME: STEP = 0.02 , TOTAL = 433.2 SECONDS ( 7.2 MIN)

CPU UTILIZATION: STEP = 86.27%, TOTAL = 99.51%

STOP IN ABRT



And this is the input file, I run a pm3 level calculation to test it.


$CONTRL RUNTYP=rsurface NZVAR=75 COORD=ZMT maxit=200 $END
$BASIS GBASIS=pm3 $END
$statpt opttol=1d-5 nstep=1000 IFREEZ(1)=48 $end
$surf ndisp1=75 disp1=5.00 vect1(48)=1 orig1=0 $end
$DATA
pes dicetona giro de los CO
c1
N
N 1 1.2345761
C 2 1.4327546 1 116.2298033
C 3 1.3813768 2 115.8067684 1 -177.2214548 0
C 4 1.3831171 3 120.0580722 2 -179.8513476 0
C 5 1.3830313 4 119.7301631 3 0.1933912 0
C 6 1.3874274 5 120.0769796 4 -0.0683608 0
C 7 1.3794521 6 120.2166425 5 -0.0221649 0
H 4 1.0702620 3 118.6378975 5 -179.8907274 0
H 5 1.0713188 4 119.9674880 6 179.9540265 0
H 6 1.0718191 5 119.9901307 7 -179.9935201 0
H 7 1.0714661 6 120.0673356 5 179.9656996 0
H 8 1.0688013 7 121.3835059 6 179.9144065 0
C 1 1.4793041 2 112.8724960 3 179.8344434 0
C 14 1.5168476 1 109.6967553 2 118.6038413 0
C 14 1.5477502 1 104.6253702 15 121.9263989 0
H 14 1.0820704 1 109.2973112 15 -121.8511526 0
O 16 1.2064300 14 118.6526461 1 00.0000000 0
O 15 1.2079787 14 122.6176240 1 27.2453158 0
C 15 1.5117095 14 114.3013769 1 -152.0700525 0
C 16 1.5079572 14 117.8881660 1 -116.3518151 0
H 21 1.0798277 16 108.8403104 14 174.6415668 0
H 21 1.0846342 16 111.2586794 22 120.5629597 0
H 21 1.0840441 16 109.6556943 22 -120.0649856 0
H 20 1.0803356 15 109.8388861 14 -164.4236754 0
H 20 1.0833438 15 111.7604994 25 122.4524896 0
H 20 1.0867870 15 108.7819477 25 -118.9581695 0
$end

Thanks at advances

Francisco Burgos


Message d4763b0bbHW-6129-558+00.htm, number 29, was posted on Thu Oct 12 '06 at 09:18:01
in reply to c8027773Sma-6129-150-00.htm

Re: PES with rsurface

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this feature is not implemented at present, however, it is easy to implement it and we'll add it into one of the forthcoming builds.

Meantime, you can split the entire scan into several smaller parts,
and this will solve your problems.

Best regards,
Alex Granovsky


On Thu Oct 12 '06, Francisco Burgos wrote
-----------------------------------------
>Hi PCGamess users.
>I am a little child using gamess and pcgamess, my question is about relaxes potential energy surfaces (PES) that is implemented in pcgamess whit rsurface. When I tried to do an scan of a torsion angle, starting on zero degrees, begin very well, but in the second step pcgamess take as input the initial coordenates, giveb ai $DATA, but with the necesary changed angle. This is not problem for the first step, but, when the running is advanced the initial coordenates are very poor and the calculations fails, at scf level or statpt level.
>Are there a way to going passing the contrained optimized structure from the step n-1 to the n-th step and so on?

>The output file is very hugh!!!! (63Mb) and at the end of this run fails with a this legende:
>
>
> MAXIMUM GRADIENT = 45.3078979 RMS GRADIENT = 7.4281827

>0OPTIMIZATION ABORTED.

> -- GRADIENT OUT OF RANGE

> -- MAXIMUM ALLOWED FORCE (FMAXT) = 10.000
>
>
>
>
>
> EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 21:46:15 LT 11-OCT-2006

> 520354 WORDS OF DYNAMIC MEMORY USED
>
>
>
> CPU TIME: STEP = 0.02 , TOTAL = 431.0 SECONDS ( 7.2 MIN)

> WALL CLOCK TIME: STEP = 0.02 , TOTAL = 433.2 SECONDS ( 7.2 MIN)

> CPU UTILIZATION: STEP = 86.27%, TOTAL = 99.51%

> STOP IN ABRT
>
>
>
>And this is the input file, I run a pm3 level calculation to test it.
>
>
> $CONTRL RUNTYP=rsurface NZVAR=75 COORD=ZMT maxit=200 $END
> $BASIS GBASIS=pm3 $END
> $statpt opttol=1d-5 nstep=1000 IFREEZ(1)=48 $end
> $surf ndisp1=75 disp1=5.00 vect1(48)=1 orig1=0 $end
> $DATA
> pes dicetona giro de los CO
>c1
> N
> N 1 1.2345761
> C 2 1.4327546 1 116.2298033
> C 3 1.3813768 2 115.8067684 1 -177.2214548 0
> C 4 1.3831171 3 120.0580722 2 -179.8513476 0
> C 5 1.3830313 4 119.7301631 3 0.1933912 0
> C 6 1.3874274 5 120.0769796 4 -0.0683608 0
> C 7 1.3794521 6 120.2166425 5 -0.0221649 0
> H 4 1.0702620 3 118.6378975 5 -179.8907274 0
> H 5 1.0713188 4 119.9674880 6 179.9540265 0
> H 6 1.0718191 5 119.9901307 7 -179.9935201 0
> H 7 1.0714661 6 120.0673356 5 179.9656996 0
> H 8 1.0688013 7 121.3835059 6 179.9144065 0
> C 1 1.4793041 2 112.8724960 3 179.8344434 0
> C 14 1.5168476 1 109.6967553 2 118.6038413 0
> C 14 1.5477502 1 104.6253702 15 121.9263989 0
> H 14 1.0820704 1 109.2973112 15 -121.8511526 0
> O 16 1.2064300 14 118.6526461 1 00.0000000 0
> O 15 1.2079787 14 122.6176240 1 27.2453158 0
> C 15 1.5117095 14 114.3013769 1 -152.0700525 0
> C 16 1.5079572 14 117.8881660 1 -116.3518151 0
> H 21 1.0798277 16 108.8403104 14 174.6415668 0
> H 21 1.0846342 16 111.2586794 22 120.5629597 0
> H 21 1.0840441 16 109.6556943 22 -120.0649856 0
> H 20 1.0803356 15 109.8388861 14 -164.4236754 0
> H 20 1.0833438 15 111.7604994 25 122.4524896 0
> H 20 1.0867870 15 108.7819477 25 -118.9581695 0
> $end

>Thanks at advances

>Francisco Burgos


Message 81313af8bJT-6129-1270-00.htm, number 30, was posted on Thu Oct 12 '06 at 21:10:13
Check examXX.inp output

Carlos A. Trallero
trallero@grad.physics.sunysb.edu


Hello all,

Is there a script for running all the exam files and check the results for PC GAMESS?

I'm thinking on something like the scripts that come with the GAMESS distribution:
runall and
checktst

I guess I could modify the scripts to run PC GAMESS on linux but has anyone done that for windows?

Thanks
Carlos


Message d4763b0bbHW-6129-1320+00.htm, number 31, was posted on Thu Oct 12 '06 at 22:00:12
in reply to 81313af8bJT-6129-1270-00.htm

Re: Check examXX.inp output

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there is no need for this at all. The PC GAMESS is distributed as the precompiled binaries which are very carefully regression tested before being uploaded to the publicly available download server.

Regards,
Alex Granovsky

On Thu Oct 12 '06, Carlos A. Trallero wrote
-------------------------------------------
>Hello all,

> Is there a script for running all the exam files and check the results for PC GAMESS?

> I'm thinking on something like the scripts that come with the GAMESS distribution:
>runall and
>checktst

> I guess I could modify the scripts to run PC GAMESS on linux but has anyone done that for windows?

> Thanks
> Carlos


Message 81313af8bJT-6136-1304-00.htm, number 32, was posted on Thu Oct 19 '06 at 21:44:01
P2P and MPI/LAM

Carlos A. Trallero
trallero@grad.physics.sunysb.edu


Dear all,

After looking at the manual pages on this web site, I can not figure out how to run pcgamess in parallel using p2p.

I know P2P has to be activated in the input file, but do I need to use mpirun to run the code? Or will PC GAMESS will run in parallel mode by it self?

Also would the load balancing work with a beowulf like queue?

Thanks
Carlos


Message d4c0f196bHW-6137-872+00.htm, number 33, was posted on Fri Oct 20 '06 at 14:31:59
in reply to 81313af8bJT-6136-1304-00.htm

Re: P2P and MPI/LAM

Alex Granovsky
gran@classic.chem.msu.su


Hi,

> After looking at the manual pages on this web site, I can not figure out how to run pcgamess in parallel using p2p.
> I know P2P has to be activated in the input file, but do I need to use mpirun to run the code? Or will PC GAMESS will run in parallel mode by it self?

P2P interface is not the replacement for MPI. You can consider MPI and P2P as two complementary interfaces which in combination allow
to perform some tasks much more efficiently than using MPI only.

As P2P initialization is performed over MPI, to use P2P you need to run PC GAMESS in parallel by the same standard way as if you were using MPI only. The only important thing here is that -ex command line option should point to the directory (the same on all nodes) which contains PC GAMESS runtime extensions (fastdiag.ex, and especially pcgp2p.ex which contains Linux-specific implementation of P2P). You also need to activate p2p interface:
$p2p p2p=.t. $end
and most likely you'll benefit from dynamic load balancing:
$p2p dlb=.t. $end
or
$p2p xdlb=.t. $end

Best regards,
Alex Granovsky


Message d4c0f196bHW-6137-937+00.htm, number 33, was edited on Fri Oct 20 '06 at 15:37:07
and replaces message d4c0f196bHW-6137-872+00.htm

Re: P2P and MPI/LAM

Alex Granovsky
gran@classic.chem.msu.su


Hi,

> After looking at the manual pages on this web site, I can not figure out how to run pcgamess in parallel using p2p.
> I know P2P has to be activated in the input file, but do I need to use mpirun to run the code? Or will PC GAMESS will run in parallel mode by it self?

P2P interface is not the replacement for MPI. You can consider MPI and P2P as two complementary interfaces which in combination allow
to perform some tasks much more efficiently than using MPI only.

As P2P initialization is performed over MPI, to use P2P you need to run PC GAMESS in parallel by the same standard way as if you were using MPI only. The only important thing here is that -ex < path > command line option should point to the directory (the same on all nodes) which contains PC GAMESS runtime extensions (fastdiag.ex, and especially pcgp2p.ex which contains Linux-specific implementation of P2P). You also need to activate p2p interface:

$p2p p2p=.t. $end

and most likely you'll benefit from dynamic load balancing:

$p2p dlb=.t. $end

or

$p2p xdlb=.t. $end

Best regards,
Alex Granovsky

[ This message was edited on Fri Oct 20 '06 by the author ]


Message c17d4ed5eOC-6140-932-00.htm, number 34, was posted on Mon Oct 23 '06 at 15:32:32
Week interactions and levels of theory

Solntsev Pasha
solntsev@univ.kiev.ua


Dear PC Gamess users.
I am going to do geometry optimization of the molecule contained copper(II) ion and two organic moieties interacting one with another. Such interaction is very week and implementation of very large basis set such as 6-31++G(3df3p) does not produce correct and accurate results. Nevertheless, introduction of the MP2 correction were shown to be able to solve this problem. Of course, I cannot use RHF function in that case; on the other hand, MP2 gradient is also unavailable under ROHF(UHF). What is the level of theory should I use due to perform molecule optimization? Are there any routine procedures for such cases? Best regards.

Message d4763b03bHW-6142-726+00.htm, number 35, was posted on Wed Oct 25 '06 at 12:06:36
in reply to c17d4ed5eOC-6140-932-00.htm

Re: Week interactions and levels of theory

Alex Granovsky
gran@classic.chem.msu.su



Hi,

I would suggest you to try to optimize your system using PBE0 functional first. PBE0 is considered to be one of the best functionals available for systems with weak interactions. If it fails, then you have no choice but to use either UMP2 or ROMP2. Unfortunately, at moment PC GAMESS lacks open-shell MP2 gradients capabilities. They will be added as a result of our current work on large-scale parallel MP2 gradient program, however, this will certainly take considerable amount of time to implement.

Best regagds,
Alex Granovsky


On Mon Oct 23 '06, Solntsev Pasha wrote
---------------------------------------
>Dear PC Gamess users.
>I am going to do geometry optimization of the molecule contained copper(II) ion and two organic moieties interacting one with another. Such interaction is very week and implementation of very large basis set such as 6-31++G(3df3p) does not produce correct and accurate results. Nevertheless, introduction of the MP2 correction were shown to be able to solve this problem. Of course, I cannot use RHF function in that case; on the other hand, MP2 gradient is also unavailable under ROHF(UHF). What is the level of theory should I use due to perform molecule optimization? Are there any routine procedures for such cases? Best regards.
>


Message c8027773Sma-6143-121-00.htm, number 36, was posted on Thu Oct 26 '06 at 02:00:57
Quiestion about PES and atomoc basis

Francisco Burgos
franciscoburgos@uach.cl


Dear PCgamess users
I am trying of perform a relaxed PES in a molecule with two transition metals (Ru and Rh) and to do this I want to give it a SBKJC basis set to the metals and 6-31g basis set to Carbons and hidrogens and I want to use ECP for metals, but I have a problem with the definition of the coordenates to scan, ( I want to separate a ligand of the metals centres). To define the diferents basis set to metals and C and H it must to be donne in Coord=unique and give the cartesian coordenates of each atoms and the basis set at the $data group. But in this way it is not possible to define an internal coordenate to perfom the r-scan job.

Am I do the right things?
Or I misunderstand the guidelines of the INPUT manual?
Are there another way to perform such job?


Message d4c0ee0abHW-6143-1297+00.htm, number 37, was posted on Thu Oct 26 '06 at 21:37:03
in reply to c8027773Sma-6143-121-00.htm

Re: Quiestion about PES and atomoc basis

Alex Granovsky
gran@classic.chem.msu.su



Hi,

you can input initial geometry using Coord=unique and still provide nonzero nzvar and valid $zmat group. The simplest way to do this is
to run check job using zmatrix input, pick up $zmat and $data
groups from the punch file, insert them into your input, and
change $data as required.

Regards,
Alex Granovsky


On Thu Oct 26 '06, Francisco Burgos wrote
-----------------------------------------
>Dear PCgamess users
>I am trying of perform a relaxed PES in a molecule with two transition metals (Ru and Rh) and to do this I want to give it a SBKJC basis set to the metals and 6-31g basis set to Carbons and hidrogens and I want to use ECP for metals, but I have a problem with the definition of the coordenates to scan, ( I want to separate a ligand of the metals centres). To define the diferents basis set to metals and C and H it must to be donne in Coord=unique and give the cartesian coordenates of each atoms and the basis set at the $data group. But in this way it is not possible to define an internal coordenate to perfom the r-scan job.

>Am I do the right things?
>Or I misunderstand the guidelines of the INPUT manual?
>Are there another way to perform such job?

>


Message c8027773Sma-6179-4-00.htm, number 38, was posted on Fri Dec 1 '06 at 00:04:37
Overlap Matrix

Francisco Burgos
franciscoburgos@uach.cl


Hi dears Pcgamess users
I have a question about the overlap matrix, Is there an option that print this overlap ?
My question is to determinate the fukui indexes directly at a single point, I am using the density of homo and lumo as an approximation, but at some articles to get a more exact value it is necesary the overlap of basis.(chem. Phys. Lett vol 304 (1999) 405.

Ok, thatks at advances

Fco. Burgos


Message d4763b0bbHW-6179-650+00.htm, number 39, was posted on Fri Dec 1 '06 at 10:50:35
in reply to c8027773Sma-6179-4-00.htm

Re: Overlap Matrix

Alex Granovsky
gran@classic.chem.msu.su


Hi,

nprint=3 will do the job. Alternatively, you can use exetyp=int1

Best regards,
Alex Granovsky

On Fri Dec 1 '06, Francisco Burgos wrote
----------------------------------------
>Hi dears Pcgamess users
>I have a question about the overlap matrix, Is there an option that print this overlap ?
>My question is to determinate the fukui indexes directly at a single point, I am using the density of homo and lumo as an approximation, but at some articles to get a more exact value it is necesary the overlap of basis.(chem. Phys. Lett vol 304 (1999) 405.

>Ok, thatks at advances

>Fco. Burgos

>


Message c8027773Sma-6189-1110-00.htm, number 40, was posted on Mon Dec 11 '06 at 18:31:10
$elden

Francisco Burgos
franciscoburgos@uach.cl


Hi dear pcgamess users
I have some question about how $elden calculate the electronic density of a molecular orbital (with morb=n n=homo or lumo).
I will like to know the formulas used by this keywords to try of get some relations between the electronic density of frontier orbitals and the fukui function.
When I use $elden morb=homo iedint=1 $end the output show the density localized on each atom and the electron density integrals.

When I use $elden morb=lumo iedint=1 $end the output show the density localized on each atom equls to zero and the electron density integrals.

My question at this point is What is the meaning of this electron density integrals? I supose that in the case of lumo the electron density is zero because the occupation of this orbital is zero.

Ok, I hope that my question dont be so basic

best regards

Francisco Burgos


Message d4c0ee0abHW-6195-1380+00.htm, number 41, was posted on Sun Dec 17 '06 at 23:00:15
in reply to c8027773Sma-6189-1110-00.htm

Re: $elden

Alex Granovsky
gran@classic.chem.msu.su



Hi,

the electron density integrals are simply the matrix elements of delta function centered at the point of interest in the AO basis, namely: (p|delta(r-r0)|q).
It is easy to see that they are simply equal to the product of values of functions p and q taken at the point r0:
(p|delta(r-r0)|q) = p(r0)*q(r0)
In the case of lumo, zero values are not related with occupation number of this MO. The density is always calculated as if this orbital has occupation number equal to one.
There are several reasons why they can be zero such as the particular nodal structure of orbital, symmetry considerations, or very low population of AOs having significant values on nuclei (e.g, diffuse orbitals).

Hope this helps.

Best regards,
Alex Granovsky


On Mon Dec 11 '06, Francisco Burgos wrote
-----------------------------------------
>Hi dear pcgamess users
>I have some question about how $elden calculate the electronic density of a molecular orbital (with morb=n n=homo or lumo).
>I will like to know the formulas used by this keywords to try of get some relations between the electronic density of frontier orbitals and the fukui function.
>When I use $elden morb=homo iedint=1 $end the output show the density localized on each atom and the electron density integrals.

>When I use $elden morb=lumo iedint=1 $end the output show the density localized on each atom equls to zero and the electron density integrals.

>My question at this point is What is the meaning of this electron density integrals? I supose that in the case of lumo the electron density is zero because the occupation of this orbital is zero.

>Ok, I hope that my question dont be so basic

>best regards

>Francisco Burgos


Message 81313af8bJT-6198-45-00.htm, number 42, was posted on Wed Dec 20 '06 at 00:44:41
PCGAMESS for Mac OS X?

Carlos A. Trallero
trallero@grad.physics.sunysb.edu


Hi Alex and PCGAMESS developers,

After trying PCGAMESS on my XP and my Linux boxes, I was wondering if there are plans to compile PCGAMESS for OS X.

Pretty sure you guys have more important things to do, but since it seems lots of people are switching to Mac nowadays it might not be too crazy.

Carlos


Message 92f4e892VQv-6198-242-00.htm, number 43, was posted on Wed Dec 20 '06 at 04:03:01
Questions about running PC GAMESS over MPI

Dong Xu
dongxu@sciences.sdsu.edu


I'm using the Serial/parallel Linux binaries version of PC GAMESS (v 7.0.1) linked with MPICH (pcgamess_linux_mpich_p4.zip).

My questions are:

1. The output file always shows "PARALLEL VERSION (MPICH) RUNNING SEQUENTIALLY ON SINGLE NODE" when running a parallel job via mpirun. Is this normal?

2. When I request a job to be run on 6 processors (-np 6) and the job is submitted to 3 nodes (2 CPUs/node, SMP) on the cluster, I only see 1 process running on each node and the second process gets lost. An the end, I got a message as follows:

p0_23745: p4_error: OOPS: semop lock failed: -1
p4_error: latest msg from perror: Identifier removed
p0_17135: p4_error: OOPS: semop lock failed: -1
p4_error: latest msg from perror: Identifier removed
p0_17856: p4_error: OOPS: semop lock failed: -1
p4_error: latest msg from perror: Identifier removed
mpiexec: Warning: tasks 0,3,5 exited with status 1.

So, I wonder if this is related to the way PC GAMESS handles SMP enviroment. Do I have to specify mklnp=2 in $SYSTEM group in order to use both processors on each node?

The command line I'm using is:

mpirun -np 6 /home/dong/bin/pcgamess_linux_mpich_p4/pcgamess -r -f -i /usr/localb/home/dong/h20/h20.inp -o /usr/localb/home/dong/h20/h20.out -ex /home/dong/
bin/pcgamess_linux_mpich_p4 -t /tmp/dong/pcgms.tmp -b /home/dong/basislib/cc-pvqz.lib


Thanks a lot!




Message 92f4e892VQv-6198-285+00.htm, number 44, was posted on Wed Dec 20 '06 at 04:45:02
in reply to d4c0f196bHW-6137-937+00.htm

Re^2: P2P and MPI/LAM

Dong Xu
dongxu@sciences.sdsu.edu


On Fri Oct 20 '06, Alex Granovsky wrote
---------------------------------------
>

Hi,

>> After looking at the manual pages on this web site, I can not figure out how to run pcgamess in parallel using p2p.
>> I know P2P has to be activated in the input file, but do I need to use mpirun to run the code? Or will PC GAMESS will run in parallel mode by it self?

>

P2P interface is not the replacement for MPI. You can consider MPI and P2P as two complementary interfaces which in combination allow
>to perform some tasks much more efficiently than using MPI only.

>

As P2P initialization is performed over MPI, to use P2P you need to run PC GAMESS in parallel by the same standard way as if you were using MPI only. The only important thing here is that -ex < path > command line option should point to the directory (the same on all nodes) which contains PC GAMESS runtime extensions (fastdiag.ex, and especially pcgp2p.ex which contains Linux-specific implementation of P2P). You also need to activate p2p interface:

Just a quick question, -ex , the path should be the directory where pcgp2p.ex can be copied from, which is usually the home directory where pcgamess binary resides. Am I right? Then -t will tell pc gamess where the pcgp2p.ex files should be copied to.


>

$p2p p2p=.t. $end

>

and most likely you'll benefit from dynamic load balancing:

>

$p2p dlb=.t. $end

>

or

>

$p2p xdlb=.t. $end

>

Best regards,
>Alex Granovsky

>


Message 92f4e892VQv-6198-287+00.htm, number 44, was edited on Wed Dec 20 '06 at 04:46:49
and replaces message 92f4e892VQv-6198-285+00.htm

Re^2: P2P and MPI/LAM

Dong Xu
dongxu@sciences.sdsu.edu


On Fri Oct 20 '06, Alex Granovsky wrote
---------------------------------------
>

Hi,

>> After looking at the manual pages on this web site, I can not figure out how to run pcgamess in parallel using p2p.
>> I know P2P has to be activated in the input file, but do I need to use mpirun to run the code? Or will PC GAMESS will run in parallel mode by it self?

>

P2P interface is not the replacement for MPI. You can consider MPI and P2P as two complementary interfaces which in combination allow
>to perform some tasks much more efficiently than using MPI only.

>

As P2P initialization is performed over MPI, to use P2P you need to run PC GAMESS in parallel by the same standard way as if you were using MPI only. The only important thing here is that -ex < path > command line option should point to the directory (the same on all nodes) which contains PC GAMESS runtime extensions (fastdiag.ex, and especially pcgp2p.ex which contains Linux-specific implementation of P2P). You also need to activate p2p interface:


>

$p2p p2p=.t. $end

>

and most likely you'll benefit from dynamic load balancing:

>

$p2p dlb=.t. $end

>

or

>

$p2p xdlb=.t. $end
-------------------------------------------------------------------

Just a quick question here, -ex , the path should be the directory where pcgp2p.ex can be copied from, which is usually the home directory where pcgamess binary resides. Then -t will tell pc gamess where the pcgp2p.ex files should be copied to. Am I right?

[ This message was edited on Wed Dec 20 '06 by the author ]


Message c8027773Sma-6198-1041+00.htm, number 45, was posted on Wed Dec 20 '06 at 17:21:49
in reply to d4c0ee0abHW-6195-1380+00.htm

Re^2: $elden

Francisco Burgos
franciscoburgos@uach.cl


Hi
Thanks for the reply.
Is it equivalent to yield the electron density of Homo (or lumo) taken the overlap matrix (given for nprint=3 in $contrl) and multiply with the coeficients matrix of the Homo (or Lumo)? and in this point take only the elemets belong to the atoms of interest???

Thaks at advances

Francisco Burgos

On Sun Dec 17 '06, Alex Granovsky wrote
---------------------------------------
>


>Hi,

>the electron density integrals are simply the matrix elements of delta function centered at the point of interest in the AO basis, namely: (p|delta(r-r0)|q).
>It is easy to see that they are simply equal to the product of values of functions p and q taken at the point r0:
>(p|delta(r-r0)|q) = p(r0)*q(r0)
>In the case of lumo, zero values are not related with occupation number of this MO. The density is always calculated as if this orbital has occupation number equal to one.
>There are several reasons why they can be zero such as the particular nodal structure of orbital, symmetry considerations, or very low population of AOs having significant values on nuclei (e.g, diffuse orbitals).

>Hope this helps.

>Best regards,
>Alex Granovsky

>


>On Mon Dec 11 '06, Francisco Burgos wrote
>-----------------------------------------
>>Hi dear pcgamess users
>>I have some question about how $elden calculate the electronic density of a molecular orbital (with morb=n n=homo or lumo).
>>I will like to know the formulas used by this keywords to try of get some relations between the electronic density of frontier orbitals and the fukui function.
>>When I use $elden morb=homo iedint=1 $end the output show the density localized on each atom and the electron density integrals.

>>When I use $elden morb=lumo iedint=1 $end the output show the density localized on each atom equls to zero and the electron density integrals.

>>My question at this point is What is the meaning of this electron density integrals? I supose that in the case of lumo the electron density is zero because the occupation of this orbital is zero.

>>Ok, I hope that my question dont be so basic

>>best regards

>>Francisco Burgos


Message 92f4f0c0VQv-6198-1356-00.htm, number 46, was posted on Wed Dec 20 '06 at 22:35:35
Spin components of E(2) in MP2 calculation?

Dong Xu
dongxu@sciences.sdsu.edu


Hi, I was wondering if there a way to output the alpha-alpha, alpha-beta, beta-beta Spin contributions to the correlation energy E(2)?

Thanks!


Message d4763b0bbHW-6199-688+00.htm, number 47, was posted on Thu Dec 21 '06 at 11:28:10
in reply to 92f4e892VQv-6198-287+00.htm

Re^3: P2P and MPI/LAM

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>Just a quick question here, -ex , the path should be the directory where pcgp2p.ex can be copied from, which is usually the home directory where pcgamess binary resides. Then -t will tell pc gamess where the pcgp2p.ex files should be copied to. Am I right?

Exactly! Note that -t switch points to directory where all the temporary files will reside, not only pcgp2p.ex.

Best regards,
Alex


Message d4763b0bbHW-6199-690+00.htm, number 48, was posted on Thu Dec 21 '06 at 11:30:40
in reply to c8027773Sma-6198-1041+00.htm

Re^3: $elden

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>Is it equivalent to yield the electron density of Homo (or lumo) taken the overlap matrix (given for nprint=3 in $contrl) and multiply with the coeficients matrix of the Homo (or Lumo)? and in this point take only the elemets belong to the atoms of interest???

No it is not.

Best regards,
Alex


Message d4763b0bbHW-6199-701+00.htm, number 49, was posted on Thu Dec 21 '06 at 11:41:15
in reply to 81313af8bJT-6198-45-00.htm

Re: PCGAMESS for Mac OS X?

Alex Granovsky
gran@classic.chem.msu.su



Hi!

At moment, we do not consider OS X too seriously. Another problem is that it is not very easy to port PC GAMESS in its current form to other OS. If there exists a version of wine for OS X, you can try to run Windows PC GAMESS version under wine. On the other hand, if there in turn will be enough systems running under OS X in the future, we'll most likely create OS X version.

Best regards,
Alex

P.S. Most of my friends who have bought iMAC are trying to replace OS X by Linux or Windows. :-)

On Wed Dec 20 '06, Carlos A. Trallero wrote
-------------------------------------------
>Hi Alex and PCGAMESS developers,

> After trying PCGAMESS on my XP and my Linux boxes, I was wondering if there are plans to compile PCGAMESS for OS X.

> Pretty sure you guys have more important things to do, but since it seems lots of people are switching to Mac nowadays it might not be too crazy.

> Carlos


Message d4763b0bbHW-6199-713+00.htm, number 50, was posted on Thu Dec 21 '06 at 11:53:25
in reply to 92f4e892VQv-6198-242-00.htm

Re: Questions about running PC GAMESS over MPI

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>1. The output file always shows "PARALLEL VERSION (MPICH) RUNNING SEQUENTIALLY ON SINGLE NODE" when running a parallel job via mpirun. Is this normal?

No this is not. This means that mpirun did not pass valid parameters to PC GAMESS master process to start parallel calculations. Check if you are using mpirun from mpich v. 1.2.x distribution, and that it is configured correctly. You can also run PC GAMESS in parallel manually by creating a proper procgrp file and using -p4pg switch (see the PC GAMESS documentation for details).

>2. When I request a job to be run on 6 processors (-np 6) and the job is submitted to 3 nodes (2 CPUs/node, SMP) on the cluster, I only see 1 process running on each node and the second process gets lost. An the end, I got a message as follows:

>p0_23745: p4_error: OOPS: semop lock failed: -1
> p4_error: latest msg from perror: Identifier removed
>p0_17135: p4_error: OOPS: semop lock failed: -1
> p4_error: latest msg from perror: Identifier removed
>p0_17856: p4_error: OOPS: semop lock failed: -1
> p4_error: latest msg from perror: Identifier removed
>mpiexec: Warning: tasks 0,3,5 exited with status 1.


Most likely this is because of:




>So, I wonder if this is related to the way PC GAMESS handles SMP enviroment. Do I have to specify mklnp=2 in $SYSTEM group in order to use both processors on each node?

No you do not.

>The command line I'm using is:

>mpirun -np 6 /home/dong/bin/pcgamess_linux_mpich_p4/pcgamess -r -f -i /usr/localb/home/dong/h20/h20.inp -o /usr/localb/home/dong/h20/h20.out -ex /home/dong/
>bin/pcgamess_linux_mpich_p4 -t /tmp/dong/pcgms.tmp -b /home/dong/basislib/cc-pvqz.lib

Looks fine. Check your mpirun/mpich version.


Best regards,
Alex





Message d4763b0bbHW-6199-723+00.htm, number 51, was posted on Thu Dec 21 '06 at 12:03:06
in reply to 92f4f0c0VQv-6198-1356-00.htm

Re: Spin components of E(2) in MP2 calculation?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

they are not calculated separately by MP2 code so you cannot get them using any currently programmed MP2 method. Ironically, MP3 code prints out singlet and triplet pairs contributions to E(2), which are directly related with contributions you are asking for.

Best regards,
Alex


On Wed Dec 20 '06, Dong Xu wrote
--------------------------------
>Hi, I was wondering if there a way to output the alpha-alpha, alpha-beta, beta-beta Spin contributions to the correlation energy E(2)?

>Thanks!


Message 44086dc4VQv-6199-1295-00.htm, number 52, was posted on Thu Dec 21 '06 at 21:35:00
"FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8" error in a MP2/cc-pVQZ job

Dong Xu
dongxu@sciences.sdsu.edu


This job is for a 26 atom large molecule. The error is under "2 ELECTRON INTEGRALS" step and it reads "FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8". What's the problem and is there a possible fix?

Thanks!


Message 57fb9196bHW-6200-816+00.htm, number 53, was posted on Fri Dec 22 '06 at 13:36:37
in reply to 44086dc4VQv-6199-1295-00.htm

Re: "FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8" error in a MP2/cc-pVQZ job

Alex Granovsky
gran@classic.chem.msu.su



Hi,

This error means that you are running out of free disk space (see FAQ section for some details).

Best regards,
Alex


On Thu Dec 21 '06, Dong Xu wrote
--------------------------------
>This job is for a 26 atom large molecule. The error is under "2 ELECTRON INTEGRALS" step and it reads "FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8". What's the problem and is there a possible fix?

>Thanks!


Message 57fb9196bHW-6200-820+00.htm, number 53, was edited on Fri Dec 22 '06 at 13:40:18
and replaces message 57fb9196bHW-6200-816+00.htm

Re: "FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8" error in a MP2/cc-pVQZ job

Alex Granovsky
gran@classic.chem.msu.su



Hi,

This error means that you are running out of free disk space (see FAQ section for some details). You can either run PC GAMESS on larger volume or switch to direct methods.

Best regards,
Alex


On Thu Dec 21 '06, Dong Xu wrote
--------------------------------
>This job is for a 26 atom large molecule. The error is under "2 ELECTRON INTEGRALS" step and it reads "FSF: fatal error no. 0xFFFFFFFB in sub PACKWRIT on unit 8". What's the problem and is there a possible fix?

>Thanks!

[ This message was edited on Fri Dec 22 '06 by the author ]


Message 44086dc4VQv-6201-1343+00.htm, number 54, was posted on Sat Dec 23 '06 at 22:23:09
in reply to d4763b0bbHW-6199-723+00.htm

Re^2: Spin components of E(2) in MP2 calculation?

Dong Xu
dongxu@sciences.sdsu.edu


Thannks! So, I just need to perform a single point MP3 energy calculation on the MP2 optimized geometry if I want to get the spin components of E(2). Correct?


On Thu Dec 21 '06, Alex Granovsky wrote
---------------------------------------
>Hi,

>they are not calculated separately by MP2 code so you cannot get them using any currently programmed MP2 method. Ironically, MP3 code prints out singlet and triplet pairs contributions to E(2), which are directly related with contributions you are asking for.

>Best regards,
>Alex
>
>
>On Wed Dec 20 '06, Dong Xu wrote
>--------------------------------
>>Hi, I was wondering if there a way to output the alpha-alpha, alpha-beta, beta-beta Spin contributions to the correlation energy E(2)?

>>Thanks!


Message 44086dc4VQv-6201-1354-00.htm, number 55, was posted on Sat Dec 23 '06 at 22:33:58
Symmetry Problem

Dong Xu
dongxu@sciences.sdsu.edu


I have an MP2/aug-cc-pVDZ job with the molecule being in C2h symmetry and it ran into this error:

MP2 TRANSFORMATION CANNOT ASSIGN A DEFINITE ORBITAL SYMMETRY TO MO 278
JOB ABORTS. YOU MUST EITHER
A) PROVIDE A SYMMETRY ADAPTED MO SET BY GUESS=MOREAD
B) ENTER NOSYM=1 TO USE THESE SYMMETRY BROKEN MO-S.

The problem is there are two 90 degree dihedral angles in this molecule (C2h symmetry) that I want to freeze throughout the calculation. If I assign NOSYM=1 to diable C2h symmstry, the optimization will turn the dihedral angles to 0 degree and find the global minimum. I didn't have this problem in my previous MP2/cc-PVDZ or MP/cc-pVTZ jobs.

So, what's the best strategy here to finish this calculation and keep the geometry staying at this saddle point (two 90 degree dihedral angles)?

Should I use RUNTYP=SADPOINT or IFREEZ/IFCART to freeze the coordinates, and how? I have COORD=CART, a full Cartisian coordinate geometry input.

Thanks in advance and Happy holidays!

-DX


Message d4763b0bbHW-6201-1385+00.htm, number 56, was posted on Sat Dec 23 '06 at 23:05:19
in reply to 44086dc4VQv-6201-1343+00.htm

Re^3: Spin components of E(2) in MP2 calculation?

Alex Granovsky
gran@classic.chem.msu.su


Well, this is certainly not the fastest way how to get them but it seems it's the only possibility at moment.

Best regards,
Alex


On Sat Dec 23 '06, Dong Xu wrote
--------------------------------
>Thannks! So, I just need to perform a single point MP3 energy calculation on the MP2 optimized geometry if I want to get the spin components of E(2). Correct?
>
>
>On Thu Dec 21 '06, Alex Granovsky wrote
>---------------------------------------
>>Hi,

>>they are not calculated separately by MP2 code so you cannot get them using any currently programmed MP2 method. Ironically, MP3 code prints out singlet and triplet pairs contributions to E(2), which are directly related with contributions you are asking for.

>>Best regards,
>>Alex
>>
>>
>>On Wed Dec 20 '06, Dong Xu wrote
>>--------------------------------
>>>Hi, I was wondering if there a way to output the alpha-alpha, alpha-beta, beta-beta Spin contributions to the correlation energy E(2)?

>>>Thanks!


Message d4763b0bbHW-6201-1394+00.htm, number 57, was posted on Sat Dec 23 '06 at 23:13:48
in reply to 44086dc4VQv-6201-1354-00.htm

Re: Symmetry Problem

Alex Granovsky
gran@classic.chem.msu.su


Are you using spherical harmonics? This problem can be due to quasi-linear dependences in the AO basis set, if not. In any case, you can try to add the following line to your input:


$moorth syms=1 symden=1 symvec=1 tolsym=1d-7 $end

to improve quality of almost everything related with orbital symmetry and to soften threshold causing PC GAMESS to abort job.

Hope this helps.

Best regards,
Alex

On Sat Dec 23 '06, Dong Xu wrote
--------------------------------
>I have an MP2/aug-cc-pVDZ job with the molecule being in C2h symmetry and it ran into this error:

> MP2 TRANSFORMATION CANNOT ASSIGN A DEFINITE ORBITAL SYMMETRY TO MO 278
> JOB ABORTS. YOU MUST EITHER
> A) PROVIDE A SYMMETRY ADAPTED MO SET BY GUESS=MOREAD
> B) ENTER NOSYM=1 TO USE THESE SYMMETRY BROKEN MO-S.

>The problem is there are two 90 degree dihedral angles in this molecule (C2h symmetry) that I want to freeze throughout the calculation. If I assign NOSYM=1 to diable C2h symmstry, the optimization will turn the dihedral angles to 0 degree and find the global minimum. I didn't have this problem in my previous MP2/cc-PVDZ or MP/cc-pVTZ jobs.

>So, what's the best strategy here to finish this calculation and keep the geometry staying at this saddle point (two 90 degree dihedral angles)?

>Should I use RUNTYP=SADPOINT or IFREEZ/IFCART to freeze the coordinates, and how? I have COORD=CART, a full Cartisian coordinate geometry input.

>Thanks in advance and Happy holidays!

>-DX


Message 44086dc4VQv-6204-1315+00.htm, number 58, was posted on Tue Dec 26 '06 at 21:55:09
in reply to d4763b0bbHW-6199-713+00.htm

Re^2: Questions about running PC GAMESS over MPI

Dong Xu
dongxu@sciences.sdsu.edu


On Thu Dec 21 '06, Alex Granovsky wrote
---------------------------------------
>Hi,

>>1. The output file always shows "PARALLEL VERSION (MPICH) RUNNING SEQUENTIALLY ON SINGLE NODE" when running a parallel job via mpirun. Is this normal?

>No this is not. This means that mpirun did not pass valid parameters to PC GAMESS master process to start parallel calculations. Check if you are using mpirun from mpich v. 1.2.x distribution, and that it is configured correctly. You can also run PC GAMESS in parallel manually by creating a proper procgrp file and using -p4pg switch (see the PC GAMESS documentation for details).

>>2. When I request a job to be run on 6 processors (-np 6) and the job is submitted to 3 nodes (2 CPUs/node, SMP) on the cluster, I only see 1 process running on each node and the second process gets lost. An the end, I got a message as follows:

>>p0_23745: p4_error: OOPS: semop lock failed: -1
>> p4_error: latest msg from perror: Identifier removed
>>p0_17135: p4_error: OOPS: semop lock failed: -1
>> p4_error: latest msg from perror: Identifier removed
>>p0_17856: p4_error: OOPS: semop lock failed: -1
>> p4_error: latest msg from perror: Identifier removed
>>mpiexec: Warning: tasks 0,3,5 exited with status 1.
>
>
>Most likely this is because of:
>


>
>
>>So, I wonder if this is related to the way PC GAMESS handles SMP enviroment. Do I have to specify mklnp=2 in $SYSTEM group in order to use both processors on each node?

>No you do not.

>>The command line I'm using is:

>>mpirun -np 6 /home/dong/bin/pcgamess_linux_mpich_p4/pcgamess -r -f -i /usr/localb/home/dong/h20/h20.inp -o /usr/localb/home/dong/h20/h20.out -ex /home/dong/
>>bin/pcgamess_linux_mpich_p4 -t /tmp/dong/pcgms.tmp -b /home/dong/basislib/cc-pvqz.lib

>Looks fine. Check your mpirun/mpich version.
>
>
>Best regards,
>Alex
>
>

The mpirun/mpich version is 1.2.6 and I think they were compiled with PGI (Portland) compilers. Could this be the problem that mpirun did not pass valid parameters to PC GAMESS master process to start parallel calculations?

Is it possible to obtain PC Gamess source and compile it locally?

Thanks!


Message d4c0ee0abHW-6204-1328+00.htm, number 59, was posted on Tue Dec 26 '06 at 22:07:56
in reply to 44086dc4VQv-6204-1315+00.htm

Re^3: Questions about running PC GAMESS over MPI

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>The mpirun/mpich version is 1.2.6 and I think they were compiled with PGI (Portland) compilers. Could this be the problem that mpirun did not pass valid parameters to PC GAMESS master process to start parallel calculations?

PC GAMESS should be compatible with your version of mpirun. Are you using some queue/batch system to run PC GAMESS? Most likely, you need to customize script which is used to run PC GAMESS to fit your needs. Note that PC GAMESS startup procedure seriously differs from that of GAMESS (US).

>Is it possible to obtain PC Gamess source and compile it locally?
We usually do not distribute the PC GAMESS sources. This will not help you in this case anyway.

Best regards,
Alex Granovsky



Message 44086dc4VQv-6204-1337+00.htm, number 60, was posted on Tue Dec 26 '06 at 22:17:38
in reply to d4c0ee0abHW-6204-1328+00.htm

Re^4: Questions about running PC GAMESS over MPI

Dong Xu
dongxu@sciences.sdsu.edu


On Tue Dec 26 '06, Alex Granovsky wrote
---------------------------------------
>Hi,

>>The mpirun/mpich version is 1.2.6 and I think they were compiled with PGI (Portland) compilers. Could this be the problem that mpirun did not pass valid parameters to PC GAMESS master process to start parallel calculations?

>PC GAMESS should be compatible with your version of mpirun. Are you using some queue/batch system to run PC GAMESS? Most likely, you need to customize script which is used to run PC GAMESS to fit your needs. Note that PC GAMESS startup procedure seriously differs from that of GAMESS (US).

>>Is it possible to obtain PC Gamess source and compile it locally?
>We usually do not distribute the PC GAMESS sources. This will not help you in this case anyway.

>Best regards,
>Alex Granovsky
>

Yes, we use PBS to submit parallel jobs and I did use correct PBS script, but the output was the same, still showing "running sequentially on a single node". The cluster is based on myrinet connection. Should I get a binary that's customized on MPICH GM?

Thank you.


Message d4c0ee0abHW-6204-1439+00.htm, number 61, was posted on Tue Dec 26 '06 at 23:58:51
in reply to 44086dc4VQv-6204-1337+00.htm

Re^5: Questions about running PC GAMESS over MPI

Alex Granovsky
gran@classic.chem.msu.su


Dear Dong,

>Yes, we use PBS to submit parallel jobs and I did use correct PBS script, but the output was the same, still showing "running sequentially on a single node". The cluster is based on myrinet connection. Should I get a binary that's customized on MPICH GM?


it seems now I understand the origin of your problems. Note, mpich and mpich-gm indeed use quite a different mechanisms to start jobs in parallel. You have two options:

Best regards,
Alex


Message 92f4e892VQv-6213-1344+00.htm, number 62, was posted on Thu Jan 4 '07 at 22:24:30
in reply to d4c0f196bHW-6000-1151+00.htm

Re^2: Parallel MP2 gradient

Dong Xu
dongxu@sciences.sdsu.edu


Prof. Granovsky,

Can you summarize the parallel capabilities in the current 7.0.2 version? I can't seem to find the info.

Thanks,

DX


On Mon Jun 5 '06, Alex Granovsky wrote
--------------------------------------
>Dear Alex,

>at present, there is no parallel MP2 gradient code in the PC GAMESS.
>Note, parallel MP2 properties are supported. We are working on
>brand new, low memory demand, efficient highly scalable
>parallel MP2 gradient code. Most likely, it will become available
>in the beta versions of the forthcoming PC GAMESS v. 7.1 and
>will be included into the final release of the PC GAMESS v. 7.1

>

>On Mon Jun 5 '06, Alex Kargovsky wrote
>--------------------------------------
>>Dear PC GAMESS users!

>>Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
>>Then I try to do these calculation type, I get this:

>>PARALLEL MP2 GRADIENT IS VERBOTEN

>>Here a my input:
>> $CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
>> $SYSTEM MWORDS=105 timlim=6000 $END
>> $scf dirscf=.t. fdiff=.f. $end
>> $STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
>> $force nvib=2 $end
>> $data
>>...

>>Thanks in advance


Message d4763b0bbHW-6214-954+00.htm, number 63, was posted on Fri Jan 5 '07 at 15:54:29
in reply to 92f4e892VQv-6213-1344+00.htm

Re^3: Parallel MP2 gradient

Alex Granovsky
gran@classic.chem.msu.su


Hi,

At present, parallel-capable code includes:

Note some solvent models are not capable to run in parallel at moment.

Multithreaded code:

see this link

Best regards,
Alex Granovsky




On Thu Jan 4 '07, Dong Xu wrote
-------------------------------
>Prof. Granovsky,

>Can you summarize the parallel capabilities in the current 7.0.2 version? I can't seem to find the info.

>Thanks,

>DX
>
>
>On Mon Jun 5 '06, Alex Granovsky wrote
>--------------------------------------
>>Dear Alex,

>>at present, there is no parallel MP2 gradient code in the PC GAMESS.
>>Note, parallel MP2 properties are supported. We are working on
>>brand new, low memory demand, efficient highly scalable
>>parallel MP2 gradient code. Most likely, it will become available
>>in the beta versions of the forthcoming PC GAMESS v. 7.1 and
>>will be included into the final release of the PC GAMESS v. 7.1

>>

>>On Mon Jun 5 '06, Alex Kargovsky wrote
>>--------------------------------------
>>>Dear PC GAMESS users!

>>>Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
>>>Then I try to do these calculation type, I get this:

>>>PARALLEL MP2 GRADIENT IS VERBOTEN

>>>Here a my input:
>>> $CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
>>> $SYSTEM MWORDS=105 timlim=6000 $END
>>> $scf dirscf=.t. fdiff=.f. $end
>>> $STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
>>> $force nvib=2 $end
>>> $data
>>>...

>>>Thanks in advance


Message 44086dc4VQv-6221-390-00.htm, number 64, was posted on Fri Jan 12 '07 at 06:30:11
mp2 full core calculation

Dong Xu
dongxu@sciences.sdsu.edu


Hi,

I think the default mp2 calculation in gamess/pc gamess is frozen core. How do I turn on full core electron correlation in mp2?

Thanks,

-DX


Message d4c0ee0abHW-6221-1377+00.htm, number 65, was posted on Fri Jan 12 '07 at 22:57:51
in reply to 44086dc4VQv-6221-390-00.htm

Re: mp2 full core calculation

Alex Granovsky
gran@classic.chem.msu.su



Hi,

just set:

$mp2 ncore=0 $end

This is described in input.doc PC GAMESS manual

Best regards,
Alex




On Fri Jan 12 '07, Dong Xu wrote
--------------------------------
>Hi,

>I think the default mp2 calculation in gamess/pc gamess is frozen core. How do I turn on full core electron correlation in mp2?

>Thanks,

>-DX


Message 44086dc4VQv-6227-1234-00.htm, number 66, was posted on Thu Jan 18 '07 at 20:34:15
Multiple-step job

Dong Xu
dongxu@sciences.sdsu.edu


How do I run a multiple-step job in one input file similar to Gaussian's Link1, the 2nd job uses the optimized geometry from the 1st job?

Thanks!


Message 92f4e892VQv-6227-1352+00.htm, number 67, was posted on Thu Jan 18 '07 at 22:32:04
in reply to d4c0ee0abHW-6221-1377+00.htm

Re^2: mp2 full core calculation

Dong Xu
dongxu@sciences.sdsu.edu


Thanks, it's actually NACORE in $MP2 group. NCORE is in $CCINP group for Coupled-cluster calculations.

On Fri Jan 12 '07, Alex Granovsky wrote
---------------------------------------
>


>Hi,

>just set:

>$mp2 ncore=0 $end

>This is described in input.doc PC GAMESS manual

>Best regards,
>Alex
>


>
>
>On Fri Jan 12 '07, Dong Xu wrote
>--------------------------------
>>Hi,

>>I think the default mp2 calculation in gamess/pc gamess is frozen core. How do I turn on full core electron correlation in mp2?

>>Thanks,

>>-DX


Message d4763b0bbHW-6228-1238+00.htm, number 68, was posted on Fri Jan 19 '07 at 20:37:53
in reply to 92f4e892VQv-6227-1352+00.htm

Re^3: mp2 full core calculation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it's NACORE in GAMESS (US) and NCORE in PC GAMESS.
Working with PC GAMESS, you should always look into the
documentation files from the PC GAMESS distribution first,
and from the GAMESS (US) second.

Regards,
Alex


On Thu Jan 18 '07, Dong Xu wrote
--------------------------------
>Thanks, it's actually NACORE in $MP2 group. NCORE is in $CCINP group for Coupled-cluster calculations.

>On Fri Jan 12 '07, Alex Granovsky wrote
>---------------------------------------
>>


>>Hi,

>>just set:

>>$mp2 ncore=0 $end

>>This is described in input.doc PC GAMESS manual

>>Best regards,
>>Alex
>>


>>
>>
>>On Fri Jan 12 '07, Dong Xu wrote
>>--------------------------------
>>>Hi,

>>>I think the default mp2 calculation in gamess/pc gamess is frozen core. How do I turn on full core electron correlation in mp2?

>>>Thanks,

>>>-DX


Message d4763b0bbHW-6228-1245+00.htm, number 69, was posted on Fri Jan 19 '07 at 20:45:02
in reply to 44086dc4VQv-6227-1234-00.htm

Re: Multiple-step job

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it is a little tricky and does not always work -
you need to have saved "dictnry" file from the first run,
then use irest=1 in $contrl, geom=daf, guess=mosaved, etc...

The much better way is to use formatted restart information
saved in "punch" file. You can create the second input file
by hand or just write a simple script to automate the process.

Regards,
Alex


On Thu Jan 18 '07, Dong Xu wrote
--------------------------------
>How do I run a multiple-step job in one input file similar to Gaussian's Link1, the 2nd job uses the optimized geometry from the 1st job?

>Thanks!


Message d4c0f196bHW-6261-862-00.htm, number 70, was posted on Wed Feb 21 '07 at 14:21:48
New PC GAMESS parallel benchmarks

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,


Press to this link to see the PC GAMESS v. 7.0.4 parallel benchmarks data on Linux Infiniband cluster


Message 9efa203aO52-6269-725-00.htm, number 71, was posted on Thu Mar 1 '07 at 12:05:07
TDDFT gradient

Khimich
khimich@kinet.chem.msu.ru


Hello Alex. Do you suggest to add TDDFT gradient in the near future? Does the beta version with available TDDFT gradient exist?

Message 9e68a878O5K-6270-1397-00.htm, number 72, was posted on Fri Mar 2 '07 at 23:39:56
Upgrade to NBO 5.x

James A Hoobler
jhoobler@willamette.edu


Hi Alex,

Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

Thanks.

James


Message 3a4467d69K2-6271-1041-00.htm, number 73, was posted on Sat Mar 3 '07 at 17:21:00
NEW USER OF PCGAMESS

neeraj misra
misraneeraj@gmail.com


I HAVE JUST DOWNLOADED PCGAMESS AND ITS .EXE BUT HOW TO START USING IT.I HAVE THE FOLLOWING QUESTIONS.
1.HOW TO MAKE INPUT FILE?
2.HOW TO RUN IT?
3.I AM INTERESTED IN OPTIMIZATION AND VIBRATIONAL FREQUENCY CALCULATIONS,WHAT SHOULD BE THE INPUT FILE STRUCTURE FOR THESE CALCULATIONS?
4.HOW TO VIEW/READ THE OUTPUT?

I SHALL BE VERY THANKFUL IF SOMEONE ANSWERS THESE QUESTIONS.


Message 57fb9096bHW-6272-859+00.htm, number 74, was posted on Sun Mar 4 '07 at 14:19:12
in reply to 9efa203aO52-6269-725-00.htm

Re: TDDFT gradient

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the TDDFT gradients will be available in the PC GAMESS version 7.2

Best regards,
Alex


On Thu Mar 1 '07, Khimich wrote
-------------------------------
>Hello Alex. Do you suggest to add TDDFT gradient in the near future? Does the beta version with available TDDFT gradient exist?


Message 57fb9096bHW-6272-866+00.htm, number 75, was posted on Sun Mar 4 '07 at 14:26:30
in reply to 9e68a878O5K-6270-1397-00.htm

Re: Upgrade to NBO 5.x

Alex Granovsky
gran@classic.chem.msu.su


Hi James,

Frankly speaking I do not know when we'll replace the NBO module. It's the question of having some spare time which is really very limited. However, there is the workaround for your problem. Namely, there exist QM/MM enabled PC GAMESS version based on the older PC GAMESS sources, which is created and maintaned by Jim Kress. Thanks to Jim, it has NBO v. 5.x incorporated into it. You can give it a try.

Best regards,
Alex


On Fri Mar 2 '07, James A Hoobler wrote
---------------------------------------
>Hi Alex,

>Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

>Thanks.

>James

>


Message 57fb9096bHW-6272-874+00.htm, number 76, was posted on Sun Mar 4 '07 at 14:34:21
in reply to 3a4467d69K2-6271-1041-00.htm

Re: NEW USER OF PCGAMESS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>1.HOW TO MAKE INPUT FILE?
Using plain text editor like notepad.
You can use built in editors of Windows Commander or FAR, etc...

>2.HOW TO RUN IT?
The simplest way is just rename it to "input",
then open command line window and run PC GAMESS.
You can use output redirection or PC GAMESS command line options
(see the command line documentation) to save output to the file.

>3.I AM INTERESTED IN OPTIMIZATION AND VIBRATIONAL FREQUENCY CALCULATIONS,WHAT SHOULD BE THE INPUT FILE STRUCTURE FOR THESE CALCULATIONS?

See the provided sample input files and read the manuals.

>4.HOW TO VIEW/READ THE OUTPUT?

Using plain text viewers or editors. Or, you can use some
visualization software like ChemCraft, etc...

Regards,
Alex


On Sat Mar 3 '07, neeraj misra wrote
------------------------------------
>I HAVE JUST DOWNLOADED PCGAMESS AND ITS .EXE BUT HOW TO START USING IT.I HAVE THE FOLLOWING QUESTIONS.
>1.HOW TO MAKE INPUT FILE?
>2.HOW TO RUN IT?
>3.I AM INTERESTED IN OPTIMIZATION AND VIBRATIONAL FREQUENCY CALCULATIONS,WHAT SHOULD BE THE INPUT FILE STRUCTURE FOR THESE CALCULATIONS?
>4.HOW TO VIEW/READ THE OUTPUT?

> I SHALL BE VERY THANKFUL IF SOMEONE ANSWERS THESE QUESTIONS.
>


Message 57fb9096bHW-6272-959+00.htm, number 75, was edited on Sun Mar 4 '07 at 15:58:44
and replaces message 57fb9096bHW-6272-866+00.htm

Re: Upgrade to NBO 5.x

Alex Granovsky
gran@classic.chem.msu.su


Hi James,

Frankly speaking I do not know when we'll replace the NBO module. It's the question of having some spare time which is really very limited. However, there is the workaround for your problem. Namely, there exist QM/MM enabled PC GAMESS version based on the older PC GAMESS sources, which was created and maintained by Jim Kress. Thanks to Jim, it has NBO v. 5.x incorporated into it. You can give it a try.

Best regards,
Alex


On Fri Mar 2 '07, James A Hoobler wrote
---------------------------------------
>Hi Alex,

>Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

>Thanks.

>James

>

[ This message was edited on Sun Mar 4 '07 by the author ]


Message d4c0f196bHW-6273-985-00.htm, number 77, was posted on Mon Mar 5 '07 at 16:25:30
JOB - Macrolanguage for PC GAMESS - is available for download

Alex Granovsky
gran@classic.chem.msu.su



Dear PC GAMESS users,

JOB - Macrolanguage for PC GAMESS (Windows), gracefully provided by Drs. Marat Talipov and Ivan Vakulin (Ufa State University) is available for download at the "Downloads" section of the PC GAMESS homepage.



Message d4c0f196bHW-6273-989-00.htm, number 78, was posted on Mon Mar 5 '07 at 16:30:06
Another PC GAMESS parallel benchmarks

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,

Press on this link to see the PC GAMESS v. 7.0.4 parallel benchmarks data on SKIF K-1000 - large Opteron Linux Infiniband cluster.


Message d4c0f196bHW-6261-862+00.htm, number 70, was edited on Mon Mar 5 '07 at 16:30:47
and replaces message d4c0f196bHW-6261-862-00.htm

New PC GAMESS parallel benchmarks

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,


Press on this link to see the PC GAMESS v. 7.0.4 parallel benchmarks data on Linux Infiniband cluster

[ This message was edited on Mon Mar 5 '07 by the author ]


Message 82e2e607UR5-6273-1085-00.htm, number 79, was posted on Mon Mar 5 '07 at 18:05:03
Simulation of reaction

greisen
pgreisen@gmail.com


Hi,

I am trying to simulate different reactions using PC-GAMESS and I am looking at the conversion of methanol to methanal. I start by building the structure of methanol and set runtyp=hessian in order to obtain a hessian for the sadpoint calculation. I take the hessian from the first run and use it in the runtyp=sadpoint.

My questions are:

1. I take the final energy from the sadpoint log-file and plot it which gives my a reverse parabola which I would expect but how to interpreted the result. Is it the reaction step
� � � � CH3OH -> CH3O- + H+

2. When I change from sadpoint to IRC I get another result where the final energy curve is a decreasing exponential which I do not expect.

Any help or advise appreciated. Thanks in advance

Best regards
Per





Message 57fb9096bHW-6275-1232+00.htm, number 75, was edited on Wed Mar 7 '07 at 20:32:30
and replaces message 57fb9096bHW-6272-959+00.htm

Re: Upgrade to NBO 5.x

Alex Granovsky
gran@classic.chem.msu.su


Hi James,

Frankly speaking I do not know when we'll replace the NBO module. It's the question of having some spare time which is really very limited. However, there is the workaround for your problem. Namely, there existsQM/MM enabled PC GAMESS version based on the older PC GAMESS sources, which was created and maintained by Jim Kress. Thanks to Jim, it has NBO v. 5.x incorporated into it. You can give it a try.

Best regards,
Alex


On Fri Mar 2 '07, James A Hoobler wrote
---------------------------------------
>Hi Alex,

>Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

>Thanks.

>James

>

[ This message was edited on Wed Mar 7 '07 by the author ]


Message 3a475164csW-6284-323-00.htm, number 80, was posted on Fri Mar 16 '07 at 05:24:02
using IFRZAT

rabin rabe
rabbipaf@yahoo.com


dear dr. granovsky,


i intend to put a diatomic oxygen on my molecule via geometry optimization. during the optimization, i freeze my molecule using the IFRZAT while the diatomic oxygen is unconstrained. i understand that IFRZAT would freeze the atoms of my molecule while it could rotate (translational, rotational) as a whole molecule. my result was not the one i was expecting. some atoms of my molecule are still moving (changing bond lengths).

my question is: is IFRZAT applicable to geometry optimization or with surface scan alone? if its not, how would i do geometry optimization by freezing my molecule and an unconstrained diatomic oxygen. hereunder is my sample input file:


$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE ECP=READ MAXIT=400
DFTTYP=B3LYP MULT=1 UNITS=ANGS irest=0 $END
$SYSTEM TIMLIM=99999 MEMORY=10000000 $END
$STATPT HESS=GUESS METHOD=QA OPTTOL=0.00001 NSTEP=999 $END


$DATA
FeC20N4H12 mult1
C1
IRON 26.0 0.0140994348 -0.0219464561 0.0279605201
SBKJC

CARBON 6.0 2.4422241720 2.4094759939 0.0387209696
N311 6

..........


H-ECP
H-ECP
O-ECP NONE
O-ECP
$END

$ZMAT DLC=.T. AUTO=.T.
IFRZAT(1)=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,
22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37
$END

any advice would be appreciated.


thanks in advance

rabin


Message 57fb9096bHW-6284-1437+00.htm, number 81, was posted on Fri Mar 16 '07 at 23:57:25
in reply to 3a475164csW-6284-323-00.htm

Re: using IFRZAT

Alex Granovsky
gran@classic.chem.msu.su


Hi Rabin,

what is the PC GAMESS version and build number you are using? Is this the official PC GAMESS v. 7.0 or 7.0.1? I'm asking as the interesting thing concerning your input is that in the release version, there is no more than 18 atoms allowed to be frozen using ifrzat, otherwise, the program will bomb. This is related with the fact that ifrzat is internally converted to constrains added to IFZMAT array, and there is no more than 50 primitive coordinates which can be frozen in the generic release version. The point is that prior to the official PC GAMESS v. 7.0 due to minor mistake in code parsing $zmat group, unrecognized keywords of $zmat group would be silently ignored. Thus, if you are using a prerelease version of the PC GAMESS v. 7.0, there can be a situation when your ifrzat group is simply ignored.

By the way, you can also use ifreez array in the $statpt group to freeze cartesian coordinates directly (although) less efficient computationally as compared with the use of ifrzat array.

Regards,
Alex Granovsky








On Fri Mar 16 '07, rabin rabe wrote
-----------------------------------
>dear dr. granovsky,
>
>
>i intend to put a diatomic oxygen on my molecule via geometry optimization. during the optimization, i freeze my molecule using the IFRZAT while the diatomic oxygen is unconstrained. i understand that IFRZAT would freeze the atoms of my molecule while it could rotate (translational, rotational) as a whole molecule. my result was not the one i was expecting. some atoms of my molecule are still moving (changing bond lengths).

>my question is: is IFRZAT applicable to geometry optimization or with surface scan alone? if its not, how would i do geometry optimization by freezing my molecule and an unconstrained diatomic oxygen. hereunder is my sample input file:
>
>
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE ECP=READ MAXIT=400
> DFTTYP=B3LYP MULT=1 UNITS=ANGS irest=0 $END
> $SYSTEM TIMLIM=99999 MEMORY=10000000 $END
> $STATPT HESS=GUESS METHOD=QA OPTTOL=0.00001 NSTEP=999 $END

>
> $DATA
> FeC20N4H12 mult1
>C1
>IRON 26.0 0.0140994348 -0.0219464561 0.0279605201
> SBKJC
>
>CARBON 6.0 2.4422241720 2.4094759939 0.0387209696
> N311 6

>..........
>
>
>H-ECP
>H-ECP
>O-ECP NONE
>O-ECP
> $END

> $ZMAT DLC=.T. AUTO=.T.
> IFRZAT(1)=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,
> 22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37
> $END

>any advice would be appreciated.
>
>
>thanks in advance

>rabin


Message 57fb9096bHW-6285-4+00.htm, number 82, was posted on Sat Mar 17 '07 at 00:03:41
in reply to 82e2e607UR5-6273-1085-00.htm

Re: Simulation of reaction

Alex Granovsky
gran@classic.chem.msu.su


Hi folks,

there are about one hundred of people already registered on this forum. While it is natural to address technical questions to the PC GAMESS developers, we would be very glad if other people will help us to maintain this forum and answer a generic questions not related with the specific knowledge of the PC GAMESS internals!

Best regards,
Alex Granovsky


On Mon Mar 5 '07, greisen wrote
-------------------------------
>Hi,

>I am trying to simulate different reactions using PC-GAMESS and I am looking at the conversion of methanol to methanal. I start by building the structure of methanol and set runtyp=hessian in order to obtain a hessian for the sadpoint calculation. I take the hessian from the first run and use it in the runtyp=sadpoint.

>My questions are:

>1. I take the final energy from the sadpoint log-file and plot it which gives my a reverse parabola which I would expect but how to interpreted the result. Is it the reaction step
> � � � � CH3OH -> CH3O- + H+

>2. When I change from sadpoint to IRC I get another result where the final energy curve is a decreasing exponential which I do not expect.

>Any help or advise appreciated. Thanks in advance

>Best regards
>Per
>
>
>
>
>


Message 3a475164csW-6287-468+00.htm, number 83, was posted on Mon Mar 19 '07 at 07:48:24
in reply to 57fb9096bHW-6284-1437+00.htm

Re^2: using IFRZAT

rabin rabe
rabbipaf@yahoo.com


hello Alex,

my downloaded binary is pcg701. however, in my output file it's PC gamess v 7.0 build # 3970. is this the one you're referring to?


rabin

On Fri Mar 16 '07, Alex Granovsky wrote
---------------------------------------
>Hi Rabin,

>what is the PC GAMESS version and build number you are using? Is this the official PC GAMESS v. 7.0 or 7.0.1? I'm asking as the interesting thing concerning your input is that in the release version, there is no more than 18 atoms allowed to be frozen using ifrzat, otherwise, the program will bomb. This is related with the fact that ifrzat is internally converted to constrains added to IFZMAT array, and there is no more than 50 primitive coordinates which can be frozen in the generic release version. The point is that prior to the official PC GAMESS v. 7.0 due to minor mistake in code parsing $zmat group, unrecognized keywords of $zmat group would be silently ignored. Thus, if you are using a prerelease version of the PC GAMESS v. 7.0, there can be a situation when your ifrzat group is simply ignored.

>By the way, you can also use ifreez array in the $statpt group to freeze cartesian coordinates directly (although) less efficient computationally as compared with the use of ifrzat array.

>Regards,
>Alex Granovsky
>
>
>
>
>
>
>
>
>On Fri Mar 16 '07, rabin rabe wrote
>-----------------------------------
>>dear dr. granovsky,
>>
>>
>>i intend to put a diatomic oxygen on my molecule via geometry optimization. during the optimization, i freeze my molecule using the IFRZAT while the diatomic oxygen is unconstrained. i understand that IFRZAT would freeze the atoms of my molecule while it could rotate (translational, rotational) as a whole molecule. my result was not the one i was expecting. some atoms of my molecule are still moving (changing bond lengths).

>>my question is: is IFRZAT applicable to geometry optimization or with surface scan alone? if its not, how would i do geometry optimization by freezing my molecule and an unconstrained diatomic oxygen. hereunder is my sample input file:
>>
>>
>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE ECP=READ MAXIT=400
>> DFTTYP=B3LYP MULT=1 UNITS=ANGS irest=0 $END
>> $SYSTEM TIMLIM=99999 MEMORY=10000000 $END
>> $STATPT HESS=GUESS METHOD=QA OPTTOL=0.00001 NSTEP=999 $END

>>
>> $DATA
>> FeC20N4H12 mult1
>>C1
>>IRON 26.0 0.0140994348 -0.0219464561 0.0279605201
>> SBKJC
>>
>>CARBON 6.0 2.4422241720 2.4094759939 0.0387209696
>> N311 6

>>..........
>>
>>
>>H-ECP
>>H-ECP
>>O-ECP NONE
>>O-ECP
>> $END

>> $ZMAT DLC=.T. AUTO=.T.
>> IFRZAT(1)=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,
>> 22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37
>> $END

>>any advice would be appreciated.
>>
>>
>>thanks in advance

>>rabin


Message d4c0f196bHW-6291-788+00.htm, number 84, was posted on Fri Mar 23 '07 at 13:08:05
in reply to 3a475164csW-6287-468+00.htm

Re^3: using IFRZAT

Alex Granovsky
gran@classic.chem.msu.su


Hi,


>my downloaded binary is pcg701. however, in my output file it's PC gamess v 7.0 build # 3970. is this the one you're referring to?

Yes it is. All PC GAMESS versions 7.0.x are still PC GAMESS
version 7.0, so that headers of outputs differ only by build #

Best regards,
Alex


Message 8afd7c53aKC-6314-934-00.htm, number 85, was posted on Sun Apr 15 '07 at 15:34:01
DLC and nonvdw

Richard
bonarlaw@liv.ac.uk


Is it possible to alter some tolerance parameter so that atoms just outside normal bonding distance get automatically picked up by auto DLC? (DLCTOL doesn't seem to help here).

Thanks

Richard


Message d4763b0bbHW-6316-1222+00.htm, number 86, was posted on Tue Apr 17 '07 at 20:22:10
in reply to 8afd7c53aKC-6314-934-00.htm

Re: DLC and nonvdw

Alex Granovsky
gran@classic.chem.msu.su


Hello,

unfortunately, this feature is not implemented at present.
You need to use nonvdw array to explicitly specify these atoms.


Best regards,
Alex Granovsky


On Sun Apr 15 '07, Richard wrote
--------------------------------
>Is it possible to alter some tolerance parameter so that atoms just outside normal bonding distance get automatically picked up by auto DLC? (DLCTOL doesn't seem to help here).

>Thanks

>Richard


Message 55ebf88bUR5-6333-61-00.htm, number 87, was posted on Fri May 4 '07 at 01:00:53
ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

greisen
pgreisen@gmail.com


Hi,

I am simulating a relatively big system - 141 atoms. I have the following input file
$CONTRL SCFTYP=RHF RUNTYP=HESSIAN MAXIT=100 ICHARG=0 MULT=1 $END
$SYSTEM TIMLIM=525600 MWORDS=76 $END
$BASIS GBASIS=N21 NGAUSS=3 $END
$SCF DIRSCF=.TRUE. DIIS=.TRUE. DAMP=.TRUE. $END
$GUESS GUESS=MOREAD $END

but whenever I execute the command I get the following error

OPENING FILE DASORT WITH 183322 LOGICAL RECORDS OF 310078 WORDS
DASORT WILL CONTAIN A MAXIMUM OF 183322 PHYSICAL RECORDS OF LENGTH 310080 WORDS
DIRECT TRANSFORMATION SKIPS AO INTEGRAL STORAGE ON DISK.
***** ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.
76138666 WORDS REQUIRED 76000000 AVAILABLE.


I try to change the mwords but it still terminates with the this error.

Any help or advise appreciated

Thanks in advance

Best regards
Per


Message d4763b0bbHW-6334-850+00.htm, number 88, was posted on Sat May 5 '07 at 14:10:31
in reply to 55ebf88bUR5-6333-61-00.htm

Re: ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

Alex Granovsky
gran@classic.chem.msu.su


Hi,

use exetyp=check to find out how much memory your run actually needs.

Regards,
Alex Granovsky


On Fri May 4 '07, greisen wrote
-------------------------------
>Hi,

>I am simulating a relatively big system - 141 atoms. I have the following input file
> $CONTRL SCFTYP=RHF RUNTYP=HESSIAN MAXIT=100 ICHARG=0 MULT=1 $END
> $SYSTEM TIMLIM=525600 MWORDS=76 $END
> $BASIS GBASIS=N21 NGAUSS=3 $END
> $SCF DIRSCF=.TRUE. DIIS=.TRUE. DAMP=.TRUE. $END
> $GUESS GUESS=MOREAD $END

>but whenever I execute the command I get the following error

> OPENING FILE DASORT WITH 183322 LOGICAL RECORDS OF 310078 WORDS
> DASORT WILL CONTAIN A MAXIMUM OF 183322 PHYSICAL RECORDS OF LENGTH 310080 WORDS
> DIRECT TRANSFORMATION SKIPS AO INTEGRAL STORAGE ON DISK.
>***** ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.
> 76138666 WORDS REQUIRED 76000000 AVAILABLE.
>
>
>I try to change the mwords but it still terminates with the this error.

>Any help or advise appreciated

>Thanks in advance

>Best regards
>Per


Message 82e157c5UR5-6335-822+00.htm, number 89, was posted on Sun May 6 '07 at 13:41:58
in reply to d4763b0bbHW-6334-850+00.htm

Re^2: ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

greisen
pgreisen@gmail.com


Hi,

I set the exetyp=check and I set the $system to

$SYSTEM TIMLIM=6000000 MWORDS=336.25 $END

which gives me the following output

TWO ELECTRON INTEGRAL DERIVATIVE CONTRIBUTIONS
----------------------------------------------

127428883 WORDS REQUIRED, 335989374 WORDS AVAILABLE
...... END OF 2-ELECTRON GRAD+HESS+FOCKDER ......


and

COUPLED-PERTURBED CLOSED SHELL HARTREE-FOCK
-------------------------------------------
THE CPHF HAS 131472 INDEPENDENT ORBITAL ROTATIONS.

*** NOT ENOUGH MEMORY TO RUN THE OUT OF MEMORY CPHF CODE ***
NWIDTH= -1 NSLICE= 131473 NAT= 143
NFREE= 53084919 NNEED= 282904455 NROT= 131472

INCREASE MEMORY IN $SYSTEM TO AT LEAST 323492695
LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!


EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 9:34:49 LT 6-MAY-2007
224425052 WORDS OF DYNAMIC MEMORY USED


which I find strange as in the two integral 335989374 WORDS AVAILABLE are available but it still asks me to increase the memory to INCREASE MEMORY IN $SYSTEM TO AT LEAST 323492695 LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!

SO what to do? Need to get more RAM on the computers? Any help or advise appreciated. Thanks in advance


Message d42eecf2bHW-6335-1173+00.htm, number 90, was posted on Sun May 6 '07 at 19:33:30
in reply to 82e157c5UR5-6335-822+00.htm

Re^3: ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please send me your compressed input file.


Regards,
Alex Granovsky


On Sun May 6 '07, greisen wrote
-------------------------------
>Hi,

>I set the exetyp=check and I set the $system to

> $SYSTEM TIMLIM=6000000 MWORDS=336.25 $END

>which gives me the following output

> TWO ELECTRON INTEGRAL DERIVATIVE CONTRIBUTIONS
> ----------------------------------------------

> 127428883 WORDS REQUIRED, 335989374 WORDS AVAILABLE
> ...... END OF 2-ELECTRON GRAD+HESS+FOCKDER ......
>
>
>and

> COUPLED-PERTURBED CLOSED SHELL HARTREE-FOCK
> -------------------------------------------
> THE CPHF HAS 131472 INDEPENDENT ORBITAL ROTATIONS.

> *** NOT ENOUGH MEMORY TO RUN THE OUT OF MEMORY CPHF CODE ***
> NWIDTH= -1 NSLICE= 131473 NAT= 143
> NFREE= 53084919 NNEED= 282904455 NROT= 131472

> INCREASE MEMORY IN $SYSTEM TO AT LEAST 323492695
> LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!
>
>
> EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 9:34:49 LT 6-MAY-2007
> 224425052 WORDS OF DYNAMIC MEMORY USED
>
>
>which I find strange as in the two integral 335989374 WORDS AVAILABLE are available but it still asks me to increase the memory to INCREASE MEMORY IN $SYSTEM TO AT LEAST 323492695 LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!

>SO what to do? Need to get more RAM on the computers? Any help or advise appreciated. Thanks in advance


Message 589de52ebQn-6337-893-00.htm, number 91, was posted on Tue May 8 '07 at 14:53:18
TDDFT with or without TDA

Pedro Silva
pedros@ufp.pt


Dear friends,

When performing a TDDFT calculation on O2 (with 6 core orbitals) I get the following error:

UNIT VECTOR GUESS AT TDDFT COEFICIENTS ...
Error: non positive-defined subspace A-B matrix!

The error does not occur if I select TDA=.true.

Why does this happen?

Pedro S.


Message a8d7d8e249G-6338-165-00.htm, number 92, was posted on Wed May 9 '07 at 02:44:44
Need more spheres for PCM

Matthew
mlardy@gmail.com


Hi,

I am having an issue with PCM in PC-Gamess. The compound that I am looking at in PCM is a little large and PC-Gamess errors letting me know that I need to increase the number of spheres. It is easy enough to do in US-Gamess (adding MXSP to the PCM line), but can I do it in PC-Gamess? I didn't see a way to do it from the available documentation, but I may have missed it.

Thanks!
Matthew


Message a8d7d8e249G-6338-167-00.htm, number 93, was posted on Wed May 9 '07 at 02:47:12
Cosmo in PC-Gamess

Matthew
mlardy@gmail.com


Hi,

Just a random question, has anyone run a Cosmo calculation (for solvation) in PC-Gamess? I haven't had much luck, and was curious if someone had been successful. (If they have been, I'll keep at it).

Thanks in advance!
Matthew


Message 82e157c5UR5-6339-841+00.htm, number 94, was posted on Thu May 10 '07 at 14:01:30
in reply to d42eecf2bHW-6335-1173+00.htm

Re^4: ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

greisen
pgreisen@gmail.com


Hi,

I have implemented the new version of pcg707_linux_mpich_p2.rar which was made available. I unpack it and replace the old files but I still get the same error when I have exetyp=check

131243138 WORDS REQUIRED, 185989374 WORDS AVAILABLE
...... END OF 2-ELECTRON GRAD+HESS+FOCKDER ......

CPU TIME: STEP = 0.62 , TOTAL = 33.1 SECONDS ( 0.6 MIN)
WALL CLOCK TIME: STEP = 4.44 , TOTAL = 76.7 SECONDS ( 1.3 MIN)
CPU UTILIZATION: STEP = 13.97%, TOTAL = 43.18%

-------------------------------------------
COUPLED-PERTURBED CLOSED SHELL HARTREE-FOCK
-------------------------------------------
THE CPHF HAS 142852 INDEPENDENT ORBITAL ROTATIONS.

*** NOT ENOUGH MEMORY TO RUN THE OUT OF MEMORY CPHF CODE ***
NWIDTH= -774 NSLICE= -1 NAT= 138
NFREE=-110648711 NNEED= 296638085 NROT= 142852

INCREASE MEMORY IN $SYSTEM TO AT LEAST 346184593
LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!

If I convert this I would need around 4 Gb of RAM to perform this calculation - is it not possible to decrease it?

Any advise or help appreciated. Thanks in advance.

Best regards


Message c1891820bQn-6339-1116+00.htm, number 95, was posted on Thu May 10 '07 at 18:36:34
in reply to a8d7d8e249G-6338-165-00.htm

Re: Need more spheres for PCM

Pedro Silva
pedros@ufp.pt


I usually take care of that problem by increasing the radius of the spheres with ret=1.0 or ret=1.5



On Wed May 9 '07, Matthew wrote
-------------------------------
>Hi,

>I am having an issue with PCM in PC-Gamess. The compound that I am looking at in PCM is a little large and PC-Gamess errors letting me know that I need to increase the number of spheres. It is easy enough to do in US-Gamess (adding MXSP to the PCM line), but can I do it in PC-Gamess? I didn't see a way to do it from the available documentation, but I may have missed it.

>Thanks!
>Matthew


Message d4763b0bbHW-6339-1132+00.htm, number 96, was posted on Thu May 10 '07 at 18:52:49
in reply to a8d7d8e249G-6338-165-00.htm

Re: Need more spheres for PCM

Alex Granovsky
gran@classic.chem.msu.su


You need to upgrade to the latest PC GAMESS builds.

Best regards,
Alex Granovsky


On Wed May 9 '07, Matthew wrote
-------------------------------
>Hi,

>I am having an issue with PCM in PC-Gamess. The compound that I am looking at in PCM is a little large and PC-Gamess errors letting me know that I need to increase the number of spheres. It is easy enough to do in US-Gamess (adding MXSP to the PCM line), but can I do it in PC-Gamess? I didn't see a way to do it from the available documentation, but I may have missed it.

>Thanks!
>Matthew


Message d4763b0bbHW-6339-1142+00.htm, number 97, was posted on Thu May 10 '07 at 19:01:55
in reply to 589de52ebQn-6337-893-00.htm

Re: TDDFT with or without TDA

Alex Granovsky
gran@classic.chem.msu.su


Hi Pedro,

this message usually related with triplet instability of the reference wavefunction. In this case, due to the nature of TDDFT equations, the transition energies become purely imaginary. However, in some rare cases this can be simply an artefact of numerical scheme used to solve TDDFT equations. As your system is small, try to increase the number of initial guess vectors or to use incore diagonalization - this should help in the latter case.

In the case of TDA/TDDFT, the equations are different and always result in real energies. In the worst case you'll simply get a negative excitation energy.

Best regards,
Alex


On Tue May 8 '07, Pedro Silva wrote
-----------------------------------
>Dear friends,

>When performing a TDDFT calculation on O2 (with 6 core orbitals) I get the following error:

> UNIT VECTOR GUESS AT TDDFT COEFICIENTS ...
> Error: non positive-defined subspace A-B matrix!

>The error does not occur if I select TDA=.true.

>Why does this happen?

>Pedro S.
>


Message d4763b0bbHW-6339-1143+00.htm, number 98, was posted on Thu May 10 '07 at 19:03:02
in reply to a8d7d8e249G-6338-167-00.htm

Re: Cosmo in PC-Gamess

Alex Granovsky
gran@classic.chem.msu.su


Hi,

At moment, COSMO is not supported.

Best regards,
Alex Granovsky

On Wed May 9 '07, Matthew wrote
-------------------------------
>Hi,

>Just a random question, has anyone run a Cosmo calculation (for solvation) in PC-Gamess? I haven't had much luck, and was curious if someone had been successful. (If they have been, I'll keep at it).

>Thanks in advance!
>Matthew


Message d4763b0bbHW-6339-1146+00.htm, number 99, was posted on Thu May 10 '07 at 19:06:11
in reply to 82e157c5UR5-6339-841+00.htm

Re^5: ERROR: MEMORY REQUEST EXCEEDS MAXIMUM.

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this means that your run does require about 347MW of memory (ca. 2.8 GB of RAM). It also requires lots of disk storage. You may prefer to switch to numerical hessian calculations.

Regards,
Alex Granovsky


On Thu May 10 '07, greisen wrote
--------------------------------
>Hi,

>I have implemented the new version of pcg707_linux_mpich_p2.rar which was made available. I unpack it and replace the old files but I still get the same error when I have exetyp=check

> 131243138 WORDS REQUIRED, 185989374 WORDS AVAILABLE
> ...... END OF 2-ELECTRON GRAD+HESS+FOCKDER ......

> CPU TIME: STEP = 0.62 , TOTAL = 33.1 SECONDS ( 0.6 MIN)
> WALL CLOCK TIME: STEP = 4.44 , TOTAL = 76.7 SECONDS ( 1.3 MIN)
> CPU UTILIZATION: STEP = 13.97%, TOTAL = 43.18%

> -------------------------------------------
> COUPLED-PERTURBED CLOSED SHELL HARTREE-FOCK
> -------------------------------------------
> THE CPHF HAS 142852 INDEPENDENT ORBITAL ROTATIONS.

> *** NOT ENOUGH MEMORY TO RUN THE OUT OF MEMORY CPHF CODE ***
> NWIDTH= -774 NSLICE= -1 NAT= 138
> NFREE=-110648711 NNEED= 296638085 NROT= 142852

> INCREASE MEMORY IN $SYSTEM TO AT LEAST 346184593
> LARGER AMOUNTS OF MEMORY ARE EVEN BETTER!

>If I convert this I would need around 4 Gb of RAM to perform this calculation - is it not possible to decrease it?

>Any advise or help appreciated. Thanks in advance.

>Best regards
>


Message d4763b0bbHW-6340-846+00.htm, number 97, was edited on Fri May 11 '07 at 14:06:18
and replaces message d4763b0bbHW-6339-1142+00.htm

Re: TDDFT with or without TDA

Alex Granovsky
gran@classic.chem.msu.su


Hi Pedro,

this message is usually related with triplet instability of the reference wavefunction. In this case, due to the nature of TDDFT equations, the transition energies become purely imaginary. However, in some rare cases this can be simply an artefact of numerical scheme used to solve TDDFT equations. As your system is small, try to increase the number of initial guess vectors or to use incore diagonalization - this should help in the latter case.

In the case of TDA/TDDFT, the equations are different and always result in real energies. In the worst case you'll simply get a negative excitation energy.

Best regards,
Alex


On Tue May 8 '07, Pedro Silva wrote
-----------------------------------
>Dear friends,

>When performing a TDDFT calculation on O2 (with 6 core orbitals) I get the following error:

> UNIT VECTOR GUESS AT TDDFT COEFICIENTS ...
> Error: non positive-defined subspace A-B matrix!

>The error does not occur if I select TDA=.true.

>Why does this happen?

>Pedro S.
>

[ This message was edited on Fri May 11 '07 by the author ]


Message 8afd7fc1aKC-6343-1122-00.htm, number 100, was posted on Mon May 14 '07 at 18:41:50
DFT settings for reference calculations

Richard
bonarlaw@liv.ac.uk


Can you recommend typical DFT settings corresponding to a general increase in precision - something equivalent to grid=fine, along with values for other tolerances mentioned in readme.dft that are likely to make any difference.

This is mainly to confirm that the default values are adequate for our particular compounds - I expect they are, but this should really be checked.

Thanks

Richard


Message d4c0ee0abHW-6343-1332+00.htm, number 101, was posted on Mon May 14 '07 at 22:12:36
in reply to 8afd7fc1aKC-6343-1122-00.htm

Re: DFT settings for reference calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I would suggest the following set of options:

$dft
NRAD=128
LMAX=47
ANGPRN=.f.
RADPRN=.f.
CUTOFF=1d-13
CUTAO=1d-13
CUTORB=1d-20
$end

Best regards,
Alex Granovsky


On Mon May 14 '07, Richard wrote
--------------------------------
>Can you recommend typical DFT settings corresponding to a general increase in precision - something equivalent to grid=fine, along with values for other tolerances mentioned in readme.dft that are likely to make any difference.

>This is mainly to confirm that the default values are adequate for our particular compounds - I expect they are, but this should really be checked.

>Thanks

>Richard


Message c22c517caGH-6344-1404-00.htm, number 102, was posted on Tue May 15 '07 at 23:24:19
PC GAMESS goes into idle state

Roman Zubatyuk
zubatyuk@gmail.com


Dear PC GAMESS developers,

I'm running PC GAMESS 7.0.7 (MPICH Pentium4 version, dynamically linked) on RedHat 9.0 system. Some runs are OK, but sometimes PC GAMESS executable goes into idle state, without any error messages.
Sometimes this happens during parallel runs. In this case PC GAMESS on one node is in idle state, but on other is 'running' with 25% user, 75% system cpu times.

Here is the report:

roman@t28 /tmp/condor/dir_23861$ tail *.log
CPU TIME: STEP = 0.00 , TOTAL = 4.5 SECONDS ( 0.1 MIN)
WALL CLOCK TIME: STEP = 0.00 , TOTAL = 4.5 SECONDS ( 0.1 MIN)
CPU UTILIZATION: STEP = 0.00%, TOTAL = 100.00%

DIRECT SCF CALCULATION, SCHWRZ=T FDIFF=F

NONZERO BLOCKS
ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR INTEGRALS SKIPPED
* * * INITIATING DIIS PROCEDURE * * *
1 0 0 -7.271573315 -7.271573315 0.974785933 0.044630982 -1 387111604
roman@t28 /tmp/condor/dir_23861$ ps xuww |grep pgamess
roman 23867 10.8 8.1 1018144 82624 ? SN 20:53 8:01 pgamess -p -i /tmp/condor/dir_23861/kamal68-cliw1-bs2.inp -o /tmp/condor/dir_23861/kamal68-cliw1-bs2.log -ex /usr/local/appls/pcgms/p4p -t /tmp/condor/dir_23861/RUN -b /usr/local/appls/pcgms/p4p
roman 23950 0.0 0.0 3572 628 pts/0 S 22:07 0:00 grep pgamess

Thanks for your help.
Roman.


Message 57fb9196bHW-6345-860+00.htm, number 103, was posted on Wed May 16 '07 at 14:20:21
in reply to c22c517caGH-6344-1404-00.htm

Re: PC GAMESS goes into idle state

Alex Granovsky
gran@classic.chem.msu.su


Dear Roman,

> I'm running PC GAMESS 7.0.7 (MPICH Pentium4 version, dynamically linked) on RedHat 9.0 system. Some runs are OK, but sometimes PC GAMESS executable goes into idle state, without any error messages.

please provide us by the the following additional information:

1. Does this happen with serial runs as well, or is only specific to parallel ones?

2. Provide us by the sample of the exact input file causing problems.

3. What is the kernel version of your Linux boxes? Does it support nptl? Are you using any multithreaded features of the PC GAMESS (async I/O, mklnp, etc...)?

4. Does the update to updated PC GAMESS v. 7.0.7 solve your problems?

Note that fastdiag.ex and pcgp2p.ex libraries from older PC GAMESS distributions are not compatible with the PC GAMESS version you are using. You must use the versions of these files from pcg707 package.

Some Linux kernels contains buggy nptl implementations. In this case, if you have linuxthreads versions of libraries in /lib/i686, try to use them via LD_LIBRARY_PATH variable.

Best regards,
Alex




>roman@t28 /tmp/condor/dir_23861$ tail *.log
> CPU TIME: STEP = 0.00 , TOTAL = 4.5 SECONDS ( 0.1 MIN)
> WALL CLOCK TIME: STEP = 0.00 , TOTAL = 4.5 SECONDS ( 0.1 MIN)
> CPU UTILIZATION: STEP = 0.00%, TOTAL = 100.00%

> DIRECT SCF CALCULATION, SCHWRZ=T FDIFF=F

> NONZERO BLOCKS
> ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR INTEGRALS SKIPPED
> * * * INITIATING DIIS PROCEDURE * * *
> 1 0 0 -7.271573315 -7.271573315 0.974785933 0.044630982 -1 387111604
>roman@t28 /tmp/condor/dir_23861$ ps xuww |grep pgamess
>roman 23867 10.8 8.1 1018144 82624 ? SN 20:53 8:01 pgamess -p -i /tmp/condor/dir_23861/kamal68-cliw1-bs2.inp -o /tmp/condor/dir_23861/kamal68-cliw1-bs2.log -ex /usr/local/appls/pcgms/p4p -t /tmp/condor/dir_23861/RUN -b /usr/local/appls/pcgms/p4p
>roman 23950 0.0 0.0 3572 628 pts/0 S 22:07 0:00 grep pgamess

> Thanks for your help.
> Roman.

>


Message 8f752c6aO8C-6346-916-00.htm, number 104, was posted on Thu May 17 '07 at 15:16:54
Second-Order MRCI

Kris
kwright06@qub.ac.uk


Hi all,

I am wanting to calculate a second order MRCI. The system has a reference CASSCF of 9 electrons in 12 orbitals resulting in more than 46000 references CSF's. Hence the total number of CSF's to use in the MRCI is over a 100 million. Is there any way of reducing this number of CSF without changing the size and occupation of the reference space?

I was thinking that by excluding those CSF's with a total contribution say less than 0.000001%, would be one solution but i'm not sure if this is possible.

Does anybody have any suggestions?

Thanks

Kris

kwright06@qub.ac.uk


Message 57fb9196bHW-6346-1032+00.htm, number 105, was posted on Thu May 17 '07 at 17:12:12
in reply to 8f752c6aO8C-6346-916-00.htm

Re: Second-Order MRCI

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you have at least three options to proceed.

First, you can try MCQDPT2 instead of MRCI. It is fast!

Second, PC GAMESS allows one to specify both SOCI and iexcit options in the $drt and $cidrt groups at the same time. In this case, iexcit refers to the maximum allowed electron excitation level.

Third, you can use $cidrt intact=.t. $end option.

Finally, you can further reduce memory needs by tuning parameters related to Davidson diagonalization in $gugdia group.

Provided you'll fit your run into available memory, I would recommend you to use pack2 option to reduce disk requirements for GUGA loops,
or even to switch to direct GUGA CI (dirci option) in the case you'll run out of free disk space.

Regards,
Alex Granovsky


On Thu May 17 '07, Kris wrote
-----------------------------
>Hi all,

>I am wanting to calculate a second order MRCI. The system has a reference CASSCF of 9 electrons in 12 orbitals resulting in more than 46000 references CSF's. Hence the total number of CSF's to use in the MRCI is over a 100 million. Is there any way of reducing this number of CSF without changing the size and occupation of the reference space?

>I was thinking that by excluding those CSF's with a total contribution say less than 0.000001%, would be one solution but i'm not sure if this is possible.

>Does anybody have any suggestions?

>Thanks

>Kris

>kwright06@qub.ac.uk


Message d4763b0bbHW-6349-1295-00.htm, number 106, was posted on Sun May 20 '07 at 21:34:40
Beta version of JOB for Linux is available for download

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,

Beta version of JOB for Linux OS is available for download at the "Downloads" section of the PC GAMESS homepage.

Best regards,
Alex Granovsky


Message ca3655832uY-6359-1150-00.htm, number 107, was posted on Wed May 30 '07 at 19:10:23
Time Dependent DFT using GAMESS

Aahana
aahanang@gmail.com


Dear GAMESS users
I am new to the use of GAMESS and I would like to to know how time dependent dft computations can be carried out using GAMESS in order to predict excited singlet and triplet states of a molecule and predict the electronic spectra.
Regards Aahana

Message d42efe42bHW-6361-1041+00.htm, number 108, was posted on Fri Jun 1 '07 at 17:21:32
in reply to ca3655832uY-6359-1150-00.htm

Re: Time Dependent DFT using GAMESS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please read the readme.tddft file from the PC GAMESS distribution.
It also can be found here

Regards,
Alex Granovsky

On Wed May 30 '07, Aahana wrote
-------------------------------
>Dear GAMESS users
>I am new to the use of GAMESS and I would like to to know how time dependent dft computations can be carried out using GAMESS in order to predict excited singlet and triplet states of a molecule and predict the electronic spectra.
>Regards Aahana


Message 806ff622euQ-6367-301-00.htm, number 109, was posted on Thu Jun 7 '07 at 05:01:22
Basic Tutorial for PC GAMESS

Kalju Kahn
kalju@chem.ucsb.edu


Hello all,

I have put together a tutorial for the basic usage of PC GAMESS. �It covers geometry optimization, frequency analysis, and the MP4 single point calculations. �I have included some thoughts on how to critically analyze the computational data. �The tutorial was written mainly with novices (undergraduate students) in mind. �Some of you might find the simple cross-platform workflow to generate input stuctures based on a SMILES-> MOL converter interesting. �

If you have the SYMMOL molecular symmetrizer installed, you can readily generate symmetric PC GAMESS input files from nearly-symmetric MOL files. �The conversion script is not thorougly tested but should work nicely in case of the Cs symmetry; manual swapping of XYZ may be needed for higher-symmetry cases. �It is not a black-box tool!

The link is posted in the Tutorial section. �I appreaciate any feedback and corrections, either via the discussion forum or privatly to kalju@chem.ucsb.edu.

Best regards,

Kalju



Message d432ac72X9o-6378-85-00.htm, number 110, was posted on Mon Jun 18 '07 at 01:25:33
GABEDIT

Christopher Wilson
chris.wilson@chemweb.org.uk


I would like to recommend gabedit as a really good gui to go with pcgamess. Definitely would be a good contender for an official gui.

Message 511794a0UVo-6386-1161-00.htm, number 111, was posted on Tue Jun 26 '07 at 19:21:21
Is there PC-GAMESS version for intel Itanium 2 CPU?

Denis Tarasov
dtarasov@mntech.ru


Is there PC-GAMESS version for intel Itanium 2 CPU? I found a web page mentioning benchmark of PC-GAMESS for Itanium but there is no official distribution on download page for this CPU. Is there any way to get Itanium version of PC-GAMESS?

Message 57fb9096bHW-6387-739+00.htm, number 112, was posted on Wed Jun 27 '07 at 12:19:12
in reply to 511794a0UVo-6386-1161-00.htm

Re: Is there PC-GAMESS version for intel Itanium 2 CPU?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Intel IA64 (Itanium/Itanium 2 CPUs) is not longer supported by PC GAMESS.

Best regards,
Alex Granovsky


On Tue Jun 26 '07, Denis Tarasov wrote
--------------------------------------
>Is there PC-GAMESS version for intel Itanium 2 CPU? I found a web page mentioning benchmark of PC-GAMESS for Itanium but there is no official distribution on download page for this CPU. Is there any way to get Itanium version of PC-GAMESS?


Message 57fb9096bHW-6387-754-00.htm, number 113, was posted on Wed Jun 27 '07 at 12:34:38
Beta test program of the PC GAMESS v. 7.1 final RC is just started.

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,

The beta test program of the PC GAMESS v. 7.1 final RC is now open.
New PC GAMESS version features completely new state of the art
parallel semidirect MP2 gradient program with excellent performance,
scalability and very modest memory demands.

The updated PC GAMESS binaries are available from MSU for registered PC GAMESS users for testing purposes upon request.

Best regards,
Alex Granovsky


Message d42c08220RG-6387-1195-00.htm, number 114, was posted on Wed Jun 27 '07 at 19:55:21
mpirun - wasn't in the Linux version (pcgamess_linux_mpich_p4.zip)

Dr Adrian Stevens
astevens@kudospharma.co.uk


Hi,

I've just downloaded the Linux version of PC Gamess (pcgamess_linux_mpich_p4.zip; linked with MPICH (5 536 578 bytes), optimized for Pentium 4, Pentium D, Xeon, and Intel Core 2 (Conroe/Merom/Woodcrest) processors.) and wanted to run some multiprocessor jobs. However, the mpirun file that is mentioned in the commandline instructions (http://classic.chem.msu.su/gran/gamess/comm_line.html) wasn't present when I unpacked the pcg70* file.

Should I have downloaded a different version of the Linux distro?

Cheers,

A.


Message d4763b0bbHW-6388-802+00.htm, number 115, was posted on Thu Jun 28 '07 at 13:21:41
in reply to d42c08220RG-6387-1195-00.htm

Re: mpirun - wasn't in the Linux version (pcgamess_linux_mpich_p4.zip)

Alex Granovsky
gran@classic.chem.msu.su


Hi,

mpirun is the standard utility from any MPI distribution.
In your particular case, you should install mpich or simply
run PC GAMESS in parallel manually by using the appropriare procgrp file.

Best regards,
Alex



On Wed Jun 27 '07, Dr Adrian Stevens wrote
------------------------------------------
>Hi,

>I've just downloaded the Linux version of PC Gamess (pcgamess_linux_mpich_p4.zip; linked with MPICH (5 536 578 bytes), optimized for Pentium 4, Pentium D, Xeon, and Intel Core 2 (Conroe/Merom/Woodcrest) processors.) and wanted to run some multiprocessor jobs. However, the mpirun file that is mentioned in the commandline instructions (http://classic.chem.msu.su/gran/gamess/comm_line.htmlhttp://classic.chem.msu.su/gran/gamess/comm_line.html) wasn't present when I unpacked the pcg70* file.

>Should I have downloaded a different version of the Linux distro?

>Cheers,

>A.


Message d4763b0bbHW-6388-803+00.htm, number 116, was posted on Thu Jun 28 '07 at 13:23:32
in reply to d4c0f196bHW-6000-1151+00.htm

Re^2: Parallel MP2 gradient

Alex Granovsky
gran@classic.chem.msu.su


Hi,

new PC GAMESS v. 7.1 has parallel MP2 gradient capability.

Best regards,
Alex



On Mon Jun 5 '06, Alex Granovsky wrote
--------------------------------------
>Dear Alex,

>at present, there is no parallel MP2 gradient code in the PC GAMESS.
>Note, parallel MP2 properties are supported. We are working on
>brand new, low memory demand, efficient highly scalable
>parallel MP2 gradient code. Most likely, it will become available
>in the beta versions of the forthcoming PC GAMESS v. 7.1 and
>will be included into the final release of the PC GAMESS v. 7.1

>

>On Mon Jun 5 '06, Alex Kargovsky wrote
>--------------------------------------
>>Dear PC GAMESS users!

>>Is it possible to perform parallel mp2 gradient calculations in PC GAMESS?
>>Then I try to do these calculation type, I get this:

>>PARALLEL MP2 GRADIENT IS VERBOTEN

>>Here a my input:
>> $CONTRL d5=.t. MAXIT=100 SCFTYP=rhf mplevl=2 RUNTYP=optimize COORD=UNIQUE $END
>> $SYSTEM MWORDS=105 timlim=6000 $END
>> $scf dirscf=.t. fdiff=.f. $end
>> $STATPT OPTTOL=0.00001 NSTEP=500 hssend=.t. $END
>> $force nvib=2 $end
>> $data
>>...

>>Thanks in advance


Message 8231aaceErv-6397-92-00.htm, number 117, was posted on Sat Jul 7 '07 at 01:32:10
How many cores?

Damodaran Krishnan
damodak@pitt.edu


On a quad-core processor, how many cores does PCGAMESS use by default?

Is there a way to specify how many cores to use?

thanks
Damodaran


Message d42eeb22bHW-6397-507+00.htm, number 118, was posted on Sat Jul 7 '07 at 08:27:30
in reply to 8231aaceErv-6397-92-00.htm

Re: How many cores?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

by default, PC GAMESS always uses only one core.
See http://classic.chem.msu.su/gran/gamess/smp.html for more information.

Best regards,
Alex Granovsky



On Sat Jul 7 '07, Damodaran Krishnan wrote
------------------------------------------
>On a quad-core processor, how many cores does PCGAMESS use by default?

>Is there a way to specify how many cores to use?

>thanks
>Damodaran


Message d43091a5T54-6400-625-00.htm, number 119, was posted on Tue Jul 10 '07 at 10:24:47
UDFT SCF convergence

Mikhail Mikhailov
mik@ioc.ac.ru


Dear PC GAMESS developers and users,

Are there any recipes (PC GAMESS specific) for improvement of UDFT SCF convergence for high spin transition metal clusters. For example, Co6 particle with spin multiplicity of 19.

Thank you in advance.
M.N. Mikhailov.


Message ca621237IDl-6407-401-00.htm, number 120, was posted on Tue Jul 17 '07 at 06:41:53
create external basis set file for , ECP external basis input

lirui
lirei1981@163.com


hi,
I can't create external basis set file for
calculation with ECP?
I can't get any working input
Thanks
lirui

Message 92a4911fXRW-6408-1430-00.htm, number 121, was posted on Wed Jul 18 '07 at 23:50:12
Problems with DFT calculations

Ricardo
teobaldo.ricardo@fis.puc-rio.br


Dear PC Gamess users
I made a DFT calculations of the protoporphyrin-Fe+3 but the results is not correct, because a energy=0, gradient=0, etc.
Thsi DFT calculations spend only 15 minutes,
Please, would you help me to find the mistake?

This is the input:
************************************************
$CONTRL RUNTYP=OPTIMIZE DFTTYP=B3LYP $END
$BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NPFUNC=1 $END
$CONTRL SCFTYP=ROHF MULT=2 $END
$CONTRL ICHARG=-1 $END
$BASIS POLAR=POPN31 $END
$DATA COORD=CART $END
$SYSTEM TIMLIM=600 MEMORY=80000000 $END
$STATPT OPTTOL=0.0001 NSTEP=1000 $END
$PDC PTSEL=CHELPG $END
$GUESS GUESS=HUCKEL $END
$DATA
Comentario
C1
C 6 -3.659291 3.681268 0.060634
C 6 -2.422115 3.002900 -0.046667
C 6 -2.696874 1.588163 -0.020966
N 7 -4.080308 1.431463 0.095164
C 6 -4.688278 2.675347 0.152980
C 6 -1.732790 0.559933 -0.099498
C 6 -1.948109 -0.794752 -0.129508
C 6 -0.900035 -1.843400 -0.193621
C 6 -1.558750 -3.060841 -0.157696
C 6 -3.009031 -2.750305 -0.094810
N 7 -3.211616 -1.423454 -0.075999
C 6 -4.027115 -3.761876 -0.069740
C 6 -5.369012 -3.530015 -0.032383
C 6 -6.454002 -4.557202 0.003288
C 6 -7.658424 -3.901778 0.104804
C 6 -7.393113 -2.435101 0.120398
N 7 -6.004632 -2.280268 -0.003177
C 6 -9.010342 -4.458119 0.175522
C 6 -8.329068 -1.453463 0.227583
C 6 -8.066172 -0.042329 0.272737
C 6 -9.159015 0.967221 0.402160
C 6 -8.551981 2.201259 0.439040
C 6 -7.077234 1.952751 0.314474
N 7 -6.866844 0.544197 0.220509
C 6 -6.251017 -5.982148 -0.054228
C 6 -5.375102 -6.604640 -0.845260
C 6 -1.000081 -4.411202 -0.182714
C 6 0.521347 -1.630726 -0.263499
C 6 1.129662 -0.708280 -1.012202
C 6 -10.583778 0.644711 0.461871
C 6 -9.205808 3.514287 0.553687
C 6 -9.991484 3.682776 1.842641
C 6 -11.492292 3.939334 1.622359
O 8 -12.139589 4.454476 2.578799
C 6 -6.099901 2.897631 0.283412
C 6 -1.095308 3.607859 -0.170090
C 6 -3.833019 5.144362 0.030644
C 6 -3.125822 5.860698 1.167020
C 6 -2.186072 6.977590 0.680917
O 8 -2.052927 7.996488 1.416513
O 8 -12.028469 3.606032 0.522658
O 8 -1.558944 6.822621 -0.411140
H 1 -6.913394 -6.565466 0.612893
H 1 -4.712892 -6.070558 -1.539482
H 1 -5.273756 -7.697164 -0.847564
H 1 -9.559610 -4.036297 1.055753
H 1 -8.983598 -5.571930 0.269522
H 1 -9.585013 -4.192691 -0.748889
H 1 -3.668393 -4.802566 -0.080811
H 1 1.118993 -2.324024 0.358009
H 1 2.221034 -0.595112 -1.018875
H 1 0.585335 -0.013079 -1.664717
H 1 -9.391759 -1.739532 0.295091
H 1 -0.684495 0.917221 -0.139623
H 1 -1.398476 -5.016208 0.671286
H 1 0.115792 -4.386192 -0.108953
H 1 -1.278630 -4.932636 -1.134053
H 1 -6.391775 3.962016 0.365047
H 1 -10.813733 0.054320 1.384589
H 1 -10.884852 0.033594 -0.425921
H 1 -11.199163 1.590511 0.478384
H 1 -0.527510 3.149275 -1.017585
H 1 -0.508275 3.449671 0.770196
H 1 -1.183834 4.720406 -0.347589
H 1 -9.927775 3.625039 -0.309833
H 1 -8.444338 4.333254 0.463932
H 1 -9.573231 4.536815 2.431835
H 1 -9.909848 2.763203 2.476362
H 1 -4.924443 5.403455 0.041121
H 1 -3.405699 5.523953 -0.944258
H 1 -2.496670 5.141980 1.752083
H 1 -3.877550 6.302533 1.867333
Fe 26 -4.919297 -0.384531 -0.009205
$END

***************************************

And this is my partial output (the last lines)


32 C 6.0 0.0000000 0.0000000 0.0000000
33 C 6.0 0.0000000 0.0000000 0.0000000
34 O 8.0 0.0000000 0.0000000 0.0000000
35 C 6.0 0.0000000 0.0000000 0.0000000
36 C 6.0 0.0000000 0.0000000 0.0000000
37 C 6.0 0.0000000 0.0000000 0.0000000
38 C 6.0 0.0000000 0.0000000 0.0000000
39 C 6.0 0.0000000 0.0000000 0.0000000
40 O 8.0 0.0000000 0.0000000 0.0000000
41 O 8.0 0.0000000 0.0000000 0.0000000
42 O 8.0 0.0000000 0.0000000 0.0000000
43 H 1.0 0.0000000 0.0000000 0.0000000
44 H 1.0 0.0000000 0.0000000 0.0000000
45 H 1.0 0.0000000 0.0000000 0.0000000
46 H 1.0 0.0000000 0.0000000 0.0000000
47 H 1.0 0.0000000 0.0000000 0.0000000
48 H 1.0 0.0000000 0.0000000 0.0000000
49 H 1.0 0.0000000 0.0000000 0.0000000
50 H 1.0 0.0000000 0.0000000 0.0000000
51 H 1.0 0.0000000 0.0000000 0.0000000
52 H 1.0 0.0000000 0.0000000 0.0000000
53 H 1.0 0.0000000 0.0000000 0.0000000
54 H 1.0 0.0000000 0.0000000 0.0000000
55 H 1.0 0.0000000 0.0000000 0.0000000
56 H 1.0 0.0000000 0.0000000 0.0000000
57 H 1.0 0.0000000 0.0000000 0.0000000
58 H 1.0 0.0000000 0.0000000 0.0000000
59 H 1.0 0.0000000 0.0000000 0.0000000
60 H 1.0 0.0000000 0.0000000 0.0000000
61 H 1.0 0.0000000 0.0000000 0.0000000
62 H 1.0 0.0000000 0.0000000 0.0000000
63 H 1.0 0.0000000 0.0000000 0.0000000
64 H 1.0 0.0000000 0.0000000 0.0000000
65 H 1.0 0.0000000 0.0000000 0.0000000
66 H 1.0 0.0000000 0.0000000 0.0000000
67 H 1.0 0.0000000 0.0000000 0.0000000
68 H 1.0 0.0000000 0.0000000 0.0000000
69 H 1.0 0.0000000 0.0000000 0.0000000
70 H 1.0 0.0000000 0.0000000 0.0000000
71 H 1.0 0.0000000 0.0000000 0.0000000
72 H 1.0 0.0000000 0.0000000 0.0000000
73 FE 26.0 0.0000000 0.0000000 0.0000000

MAXIMUM GRADIENT = 0.0000000 RMS GRADIENT = 0.0000000

***** EQUILIBRIUM GEOMETRY LOCATED *****
CHECK SHOWS $TRUNCN GROUP INPUT IS NOT PRESENT
9207959 WORDS OF DYNAMIC MEMORY USED
...etc



Sincerely yours


Ricardo


Message 4df668afbHW-6412-1334+00.htm, number 122, was posted on Sun Jul 22 '07 at 22:14:51
in reply to d43091a5T54-6400-625-00.htm

Re: UDFT SCF convergence

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you can try to change the value of fshift variable of the $scf group, which defines the constant level shift parameter used by the PC GAMESS' SCF code. The default value of 0.1 is appropriate in many cases, however, you can try to increase it to, say, 0.2 or soften to 0.05 or even 0.02, or disable level shifting completely (fshift=0.0). Typically, larger values of fshift should be used in complex cases but they generally negatively impact SCF convergence rate.

Best regards,
Alex Granovsky



On Tue Jul 10 '07, Mikhail Mikhailov wrote
------------------------------------------
>Dear PC GAMESS developers and users,

>Are there any recipes (PC GAMESS specific) for improvement of UDFT SCF convergence for high spin transition metal clusters. For example, Co6 particle with spin multiplicity of 19.

>Thank you in advance.
>M.N. Mikhailov.


Message 4df668afbHW-6412-1341+00.htm, number 123, was posted on Sun Jul 22 '07 at 22:21:16
in reply to ca621237IDl-6407-401-00.htm

Re: create external basis set file for , ECP external basis input

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. See the PC GAMESS' input preprocessing sample at this page

2. Within several days, we'll put def-2 basis sets library and script to handle it (the contribution gracefully donated by Roman Zubatyuk) to the PC GAMESS' download pages.

Best regards,
Alex Granovsky




On Tue Jul 17 '07, lirui wrote
------------------------------
>hi,
>I can't create external basis set file for
>calculation with ECP?
>I can't get any working input
>Thanks
>lirui


Message 4df668afbHW-6412-1346+00.htm, number 124, was posted on Sun Jul 22 '07 at 22:26:44
in reply to 92a4911fXRW-6408-1430-00.htm

Re: Problems with DFT calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. it seems you are using WinGamess or GAMESS (US), not the PC GAMESS (as PC GAMESS does not check for $TRUNCN group).
2. It seems there is exetyp=check directive somewhere in your input.

The following modifications allows your input to run successfully with PC GAMESS:

 $CONTRL SCFTYP=UHF MULT=2 $END
 $BASIS POLAR=POPLE $END

Best regards,
Alex Granovsky



On Wed Jul 18 '07, Ricardo wrote
--------------------------------
>Dear PC Gamess users
>I made a DFT calculations of the protoporphyrin-Fe+3 but the results is not correct, because a energy=0, gradient=0, etc.
>Thsi DFT calculations spend only 15 minutes,
>Please, would you help me to find the mistake?

>This is the input:
>************************************************
> $CONTRL RUNTYP=OPTIMIZE DFTTYP=B3LYP $END
> $BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NPFUNC=1 $END
> $CONTRL SCFTYP=ROHF MULT=2 $END
> $CONTRL ICHARG=-1 $END
> $BASIS POLAR=POPN31 $END
> $DATA COORD=CART $END
> $SYSTEM TIMLIM=600 MEMORY=80000000 $END
> $STATPT OPTTOL=0.0001 NSTEP=1000 $END
> $PDC PTSEL=CHELPG $END
> $GUESS GUESS=HUCKEL $END
> $DATA
>Comentario
>C1
> C 6 -3.659291 3.681268 0.060634
> C 6 -2.422115 3.002900 -0.046667
> C 6 -2.696874 1.588163 -0.020966
> N 7 -4.080308 1.431463 0.095164
> C 6 -4.688278 2.675347 0.152980
> C 6 -1.732790 0.559933 -0.099498
> C 6 -1.948109 -0.794752 -0.129508
> C 6 -0.900035 -1.843400 -0.193621
> C 6 -1.558750 -3.060841 -0.157696
> C 6 -3.009031 -2.750305 -0.094810
> N 7 -3.211616 -1.423454 -0.075999
> C 6 -4.027115 -3.761876 -0.069740
> C 6 -5.369012 -3.530015 -0.032383
> C 6 -6.454002 -4.557202 0.003288
> C 6 -7.658424 -3.901778 0.104804
> C 6 -7.393113 -2.435101 0.120398
> N 7 -6.004632 -2.280268 -0.003177
> C 6 -9.010342 -4.458119 0.175522
> C 6 -8.329068 -1.453463 0.227583
> C 6 -8.066172 -0.042329 0.272737
> C 6 -9.159015 0.967221 0.402160
> C 6 -8.551981 2.201259 0.439040
> C 6 -7.077234 1.952751 0.314474
> N 7 -6.866844 0.544197 0.220509
> C 6 -6.251017 -5.982148 -0.054228
> C 6 -5.375102 -6.604640 -0.845260
> C 6 -1.000081 -4.411202 -0.182714
> C 6 0.521347 -1.630726 -0.263499
> C 6 1.129662 -0.708280 -1.012202
> C 6 -10.583778 0.644711 0.461871
> C 6 -9.205808 3.514287 0.553687
> C 6 -9.991484 3.682776 1.842641
> C 6 -11.492292 3.939334 1.622359
> O 8 -12.139589 4.454476 2.578799
> C 6 -6.099901 2.897631 0.283412
> C 6 -1.095308 3.607859 -0.170090
> C 6 -3.833019 5.144362 0.030644
> C 6 -3.125822 5.860698 1.167020
> C 6 -2.186072 6.977590 0.680917
> O 8 -2.052927 7.996488 1.416513
> O 8 -12.028469 3.606032 0.522658
> O 8 -1.558944 6.822621 -0.411140
> H 1 -6.913394 -6.565466 0.612893
> H 1 -4.712892 -6.070558 -1.539482
> H 1 -5.273756 -7.697164 -0.847564
> H 1 -9.559610 -4.036297 1.055753
> H 1 -8.983598 -5.571930 0.269522
> H 1 -9.585013 -4.192691 -0.748889
> H 1 -3.668393 -4.802566 -0.080811
> H 1 1.118993 -2.324024 0.358009
> H 1 2.221034 -0.595112 -1.018875
> H 1 0.585335 -0.013079 -1.664717
> H 1 -9.391759 -1.739532 0.295091
> H 1 -0.684495 0.917221 -0.139623
> H 1 -1.398476 -5.016208 0.671286
> H 1 0.115792 -4.386192 -0.108953
> H 1 -1.278630 -4.932636 -1.134053
> H 1 -6.391775 3.962016 0.365047
> H 1 -10.813733 0.054320 1.384589
> H 1 -10.884852 0.033594 -0.425921
> H 1 -11.199163 1.590511 0.478384
> H 1 -0.527510 3.149275 -1.017585
> H 1 -0.508275 3.449671 0.770196
> H 1 -1.183834 4.720406 -0.347589
> H 1 -9.927775 3.625039 -0.309833
> H 1 -8.444338 4.333254 0.463932
> H 1 -9.573231 4.536815 2.431835
> H 1 -9.909848 2.763203 2.476362
> H 1 -4.924443 5.403455 0.041121
> H 1 -3.405699 5.523953 -0.944258
> H 1 -2.496670 5.141980 1.752083
> H 1 -3.877550 6.302533 1.867333
>Fe 26 -4.919297 -0.384531 -0.009205
>$END

>***************************************

>And this is my partial output (the last lines)
>
>
> 32 C 6.0 0.0000000 0.0000000 0.0000000
> 33 C 6.0 0.0000000 0.0000000 0.0000000
> 34 O 8.0 0.0000000 0.0000000 0.0000000
> 35 C 6.0 0.0000000 0.0000000 0.0000000
> 36 C 6.0 0.0000000 0.0000000 0.0000000
> 37 C 6.0 0.0000000 0.0000000 0.0000000
> 38 C 6.0 0.0000000 0.0000000 0.0000000
> 39 C 6.0 0.0000000 0.0000000 0.0000000
> 40 O 8.0 0.0000000 0.0000000 0.0000000
> 41 O 8.0 0.0000000 0.0000000 0.0000000
> 42 O 8.0 0.0000000 0.0000000 0.0000000
> 43 H 1.0 0.0000000 0.0000000 0.0000000
> 44 H 1.0 0.0000000 0.0000000 0.0000000
> 45 H 1.0 0.0000000 0.0000000 0.0000000
> 46 H 1.0 0.0000000 0.0000000 0.0000000
> 47 H 1.0 0.0000000 0.0000000 0.0000000
> 48 H 1.0 0.0000000 0.0000000 0.0000000
> 49 H 1.0 0.0000000 0.0000000 0.0000000
> 50 H 1.0 0.0000000 0.0000000 0.0000000
> 51 H 1.0 0.0000000 0.0000000 0.0000000
> 52 H 1.0 0.0000000 0.0000000 0.0000000
> 53 H 1.0 0.0000000 0.0000000 0.0000000
> 54 H 1.0 0.0000000 0.0000000 0.0000000
> 55 H 1.0 0.0000000 0.0000000 0.0000000
> 56 H 1.0 0.0000000 0.0000000 0.0000000
> 57 H 1.0 0.0000000 0.0000000 0.0000000
> 58 H 1.0 0.0000000 0.0000000 0.0000000
> 59 H 1.0 0.0000000 0.0000000 0.0000000
> 60 H 1.0 0.0000000 0.0000000 0.0000000
> 61 H 1.0 0.0000000 0.0000000 0.0000000
> 62 H 1.0 0.0000000 0.0000000 0.0000000
> 63 H 1.0 0.0000000 0.0000000 0.0000000
> 64 H 1.0 0.0000000 0.0000000 0.0000000
> 65 H 1.0 0.0000000 0.0000000 0.0000000
> 66 H 1.0 0.0000000 0.0000000 0.0000000
> 67 H 1.0 0.0000000 0.0000000 0.0000000
> 68 H 1.0 0.0000000 0.0000000 0.0000000
> 69 H 1.0 0.0000000 0.0000000 0.0000000
> 70 H 1.0 0.0000000 0.0000000 0.0000000
> 71 H 1.0 0.0000000 0.0000000 0.0000000
> 72 H 1.0 0.0000000 0.0000000 0.0000000
> 73 FE 26.0 0.0000000 0.0000000 0.0000000

> MAXIMUM GRADIENT = 0.0000000 RMS GRADIENT = 0.0000000

> ***** EQUILIBRIUM GEOMETRY LOCATED *****
> CHECK SHOWS $TRUNCN GROUP INPUT IS NOT PRESENT
>9207959 WORDS OF DYNAMIC MEMORY USED
>...etc
>
>
>
>Sincerely yours
>
>
>Ricardo


Message d4763b0bbHW-6423-843+00.htm, number 125, was posted on Thu Aug 2 '07 at 14:03:16
in reply to 4df668afbHW-6412-1341+00.htm

Re^2: create external basis set file for , ECP external basis input

Alex Granovsky
gran@classic.chem.msu.su


Hi,

The def2 basis set definition library as well as Python script to handle basis set/ECP definition files are now available at the "Downloads" section of the PC GAMESS homepage. Thanks to Roman Zubatyuk for his contribution!

Best regards,
Alex Granovsky


Message d4763b0bbHW-6423-1042+00.htm, number 126, was posted on Thu Aug 2 '07 at 17:22:29
in reply to d4763b0bbHW-5994-1102+00.htm

Re^2: inputs for transition metal calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

we have just added the "Lesson 2: converging difficult SCF cases" page to the PC GAMESS' Tutorials section. Hope it will be of some help for our users.

Best regards,
Alex Granovsky


Message 58d64f5eeOC-6425-87-00.htm, number 127, was posted on Sat Aug 4 '07 at 01:28:18
Version of PC GAMESS

Solntsev Pasha
solntsev@univ.kiev.ua


Hi, I get message in my output file
WARNING! YOU ARE USING OUTDATED VERSION OF THE PC GAMESS!
PLEASE CHECK PC GAMESS HOMEPAGE FOR INFORMATION ON UPDATES!
what does it mean? Many thanks.

Message 51b344b34el-6425-988-00.htm, number 128, was posted on Sat Aug 4 '07 at 16:28:32
Unable to Project DLC

Adam Sykes
adam.sykes@liv.ac.uk


Hello,

I am receiving the error message: Unable to Project DLC, when I do a check run on one of my input files. Other similar calculations work fine. Can anyone tell me what causes this and how to get around it?

Thank you,

Adam Sykes
The University of Liverpool, UK


Message 8d144c19Rck-6425-1415-00.htm, number 129, was posted on Sat Aug 4 '07 at 23:34:54
Problem with calculation of dipole transition moments

Ilya Ioffe
ioffe@thermo.chem.msu.ru


Dear Alexandr Alexandrovich,

I've encountered a strange problem while trying to calculate dipole transition moments (GUGA CI). The software ignores my input of NFZC and NOCC in the $TRANST group (other variables, such as NUMCI, are read correctly, that has been tested) and automatically sets them to the full basis set dimension. After that, the software, reports, not surprisingly, inconsistency with $CIDRT1 data and terminates. What could that be?

Sincerely yours,

Ilya Ioffe


Message 589de659bQn-6427-40+00.htm, number 130, was posted on Mon Aug 6 '07 at 00:40:14
in reply to 51b344b34el-6425-988-00.htm

Re: Unable to Project DLC

Pedro Silva
pedros@ufp.pt


Dear Adam,

In these situations, I usually define extra bonds between atoms through

$zmat nonvdw(1)=..... $end

I start by defining bonds between separate fragments, and after that, I just define random bonds. I think there is a systematic way to do it, but with some luck I often get around that problem!


Pedro


Message d4763b0bbHW-6427-817+00.htm, number 131, was posted on Mon Aug 6 '07 at 13:37:11
in reply to 8d144c19Rck-6425-1415-00.htm

Re: Problem with calculation of dipole transition moments

Alex Granovsky
gran@classic.chem.msu.su


Hi Ilya,

most likely this is some misprint in the input file.
Could you please send me the compressed input?

Best regards,
Alex Granovsky




On Sat Aug 4 '07, Ilya Ioffe wrote
----------------------------------
>Dear Alexandr Alexandrovich,

>I've encountered a strange problem while trying to calculate dipole transition moments (GUGA CI). The software ignores my input of NFZC and NOCC in the $TRANST group (other variables, such as NUMCI, are read correctly, that has been tested) and automatically sets them to the full basis set dimension. After that, the software, reports, not surprisingly, inconsistency with $CIDRT1 data and terminates. What could that be?

>Sincerely yours,

>Ilya Ioffe


Message d4763b0bbHW-6427-830+00.htm, number 132, was posted on Mon Aug 6 '07 at 13:50:28
in reply to 58d64f5eeOC-6425-87-00.htm

Re: Version of PC GAMESS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this is because of internal "soft" timebomb in the PC GAMESS executables. Initially, we planned to release PC GAMESS v. 7.1 several weeks before this message would appear. Right now it should be considered simply as a reminder that PC GAMESS v. 7.1 will be releases within a week or so and that we strongly recommend our users to pgrade to this PC GAMESS version when it will become publicly available.

Every PC GAMESS version in the future will have this time-triggered reminder build in. We hope this will make the upgrade process more regular and frequent - and this is important as we constantly add new features and bugfixes, as well as performance improvements.

Best regards,
Alex


On Sat Aug 4 '07, Solntsev Pasha wrote
--------------------------------------
>Hi, I get message in my output file
> WARNING! YOU ARE USING OUTDATED VERSION OF THE PC GAMESS!
> PLEASE CHECK PC GAMESS HOMEPAGE FOR INFORMATION ON UPDATES!
>what does it mean? Many thanks.


Message d4763b0bbHW-6427-897+00.htm, number 131, was edited on Mon Aug 6 '07 at 14:56:59
and replaces message d4763b0bbHW-6427-817+00.htm

Re: Problem with calculation of dipole transition moments

Alex Granovsky
gran@classic.chem.msu.su


Hi Ilya,

most likely this is because of some misprint in the input file.
Could you please send me the compressed input?

Best regards,
Alex Granovsky




On Sat Aug 4 '07, Ilya Ioffe wrote
----------------------------------
>Dear Alexandr Alexandrovich,

>I've encountered a strange problem while trying to calculate dipole transition moments (GUGA CI). The software ignores my input of NFZC and NOCC in the $TRANST group (other variables, such as NUMCI, are read correctly, that has been tested) and automatically sets them to the full basis set dimension. After that, the software, reports, not surprisingly, inconsistency with $CIDRT1 data and terminates. What could that be?

>Sincerely yours,

>Ilya Ioffe

[ This message was edited on Mon Aug 6 '07 by the author ]


Message d4763b0bbHW-6427-899+00.htm, number 132, was edited on Mon Aug 6 '07 at 14:59:06
and replaces message d4763b0bbHW-6427-830+00.htm

Re: Version of PC GAMESS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this is because of internal "soft" timebomb in the PC GAMESS executables. Initially, we planned to release PC GAMESS v. 7.1 several weeks before this message would appear. Right now it should be considered simply as a reminder that PC GAMESS v. 7.1 will be released within a week or so and that we strongly recommend our users to upgrade to this PC GAMESS version when it will become publicly available.

Every PC GAMESS version in the future will have this time-triggered reminder build in. We hope this will make the upgrade process more regular and frequent - and this is important as we constantly add new features and bugfixes, as well as performance improvements.

Best regards,
Alex


On Sat Aug 4 '07, Solntsev Pasha wrote
--------------------------------------
>Hi, I get message in my output file
> WARNING! YOU ARE USING OUTDATED VERSION OF THE PC GAMESS!
> PLEASE CHECK PC GAMESS HOMEPAGE FOR INFORMATION ON UPDATES!
>what does it mean? Many thanks.

[ This message was edited on Mon Aug 6 '07 by the author ]


Message 5b7cc43fPMq-6448-742-00.htm, number 133, was posted on Mon Aug 27 '07 at 12:22:14
Density of States in PC GAMESS

Vitaliy Gorbenko
vgorbenko@ukr.net


Hi,

How to find DOS from calculation results? Is there software for the visualization DOS from PC-GAMESS's output files?
For example, GaussSum is useful in the work with GAMESS output but unfortunately it don't work with PC GAMESS output.


Message d4763b0bbHW-6448-1033+00.htm, number 134, was posted on Mon Aug 27 '07 at 17:13:19
in reply to 5b7cc43fPMq-6448-742-00.htm

Re: Density of States in PC GAMESS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. AOMix
2. Latest GausSum should work. If it does not, please submit bug report to GausSum developers.

Best regards,
Alex Granovsky


On Mon Aug 27 '07, Vitaliy Gorbenko wrote
-----------------------------------------
>Hi,

>How to find DOS from calculation results? Is there software for the visualization DOS from PC-GAMESS's output files?
>For example, GaussSum is useful in the work with GAMESS output but unfortunately it don't work with PC GAMESS output.


Message be88725begc-6452-1141-00.htm, number 135, was posted on Fri Aug 31 '07 at 19:00:46
SCC-DFTB

marcos Villarreal
arloa@mail.fcq.unc.edu.ar



Dear all,
is the SCC-DFTB method implemented in PC GAMSS?

Thanks,
Marcos Villarreal


Message d4763b0bbHW-6453-1304+00.htm, number 136, was posted on Sat Sep 1 '07 at 21:44:25
in reply to be88725begc-6452-1141-00.htm

Re: SCC-DFTB

Alex Granovsky
gran@classic.chem.msu.su


Dear Marcos,

no it is not.

Regards,
Alex Granovsky



On Fri Aug 31 '07, marcos Villarreal wrote
------------------------------------------
>>Dear all,
>is the SCC-DFTB method implemented in PC GAMESS?

>Thanks,
>Marcos Villarreal


Message d4763b0bbHW-6458-922-00.htm, number 137, was posted on Thu Sep 6 '07 at 15:21:51
Release of the final PC GAMESS version 7.1

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,

the new PC GAMESS version 7.1 was released on September 5, 2007.

Please update your PC GAMESS binaries.

Best regards,
Alex Granovsky


Message cbbd308acsN-6461-515-00.htm, number 138, was posted on Sun Sep 9 '07 at 08:34:57
Linux or Windows; Xeon Quad or Dual core?

Rudy
rcoquet_mailing@fastmail.fm


Dear all,

As a new user of this forum, I would like to thank Dr A. Granovsky and his team for developing and maintaining PC Gamess.

I have some basic questions regarding the OS and CPU I should use with PC Gamess:

1. OS
Of these four OS, which one would be "the best": Windows Vista 64, Windows XP (32 or 64), Red Hat Enterprise WS V4 (64) and Red Hat Enterprise 5 Desktop (64). I am also considering Fedora 7.
I heard the parallel I/O under Windows is not very efficient...

2. CPU
I am considering a single PC with either a Xeon Quad core E5335 or two Xeon Dual core 5130. The E5335 is more or less equivalent to two 5130 put together so I wonder which solution would be the best for PC Gamess: 1CPU with 4 cores (ie. E5335) or 2 CPUs with 2 cores each (5130).

Thanks a lot for your help,
Rudy.


Message d4763b0bbHW-6463-1161+00.htm, number 139, was posted on Tue Sep 11 '07 at 19:21:31
in reply to cbbd308acsN-6461-515-00.htm

Re: Linux or Windows; Xeon Quad or Dual core?

Alex Granovsky
gran@classic.chem.msu.su


Dear Rudy,

the questions you are raising are the difficult ones. I would be glad if every users of this forum will join this discussion. I can provide you only by several hints which could make the decision simpler.

1. Definitely, the OS should be the 64-bit one. This is because PC GAMESS uses 64-bit computation kernels for better performance running under 64-bit OS. This is clearly impossible under 32-bit OS.

2. PC GAMESS runs equally fast under Linux and Windows.

>I heard the parallel I/O under Windows is not very efficient...

3. The problem is that it is not better under Linux, and is even worse in most cases.

4. Running under Linux, PC GAMESS will be able to allocate more memory.

5. Are you planning to use these systems exclusively for PC GAMESS or not? If you aren't you should consider the availability of other software for the selected OS.


>2. CPU
>I am considering a single PC with either a Xeon Quad core E5335 or two Xeon Dual core 5130. The E5335 is more or less equivalent to two 5130 put together so I wonder which solution would be the best for PC Gamess: 1CPU with 4 cores (ie. E5335) or 2 CPUs with 2 cores each (5130).

I would opt for two dual-core CPUs. The reason is that the aggregate memory bandwidth is higher in the latter case due to DIB feature of chipset. Thus, the scalability would be somewhat better. Do not forget to buy four HDDs and organize them as four independent disks, not the raid array - you will then be able to assign a separate HDD as a scratch to each instance of the parallel PC GAMESS process.


Best regards,
Alex Granovsky


On Sun Sep 9 '07, Rudy wrote
----------------------------
>Dear all,

>As a new user of this forum, I would like to thank Dr A. Granovsky and his team for developing and maintaining PC Gamess.

>I have some basic questions regarding the OS and CPU I should use with PC Gamess:

>1. OS
>Of these four OS, which one would be "the best": Windows Vista 64, Windows XP (32 or 64), Red Hat Enterprise WS V4 (64) and Red Hat Enterprise 5 Desktop (64). I am also considering Fedora 7.
>I heard the parallel I/O under Windows is not very efficient...

>2. CPU
>I am considering a single PC with either a Xeon Quad core E5335 or two Xeon Dual core 5130. The E5335 is more or less equivalent to two 5130 put together so I wonder which solution would be the best for PC Gamess: 1CPU with 4 cores (ie. E5335) or 2 CPUs with 2 cores each (5130).

>Thanks a lot for your help,
>Rudy.


Message 935b012devw-6469-935-00.htm, number 140, was posted on Mon Sep 17 '07 at 15:37:02
two quad machines cluster

Zoran Matovic
zmatovic@kg.ac.yu


Wish to ask You: I am planning to build up small cluster of two machines with quadripole processors.

Please have a look the configuration of two identical machines given bellow and help me to decide where different solutions exists:

Processor: Intel LGA Core2Quad Q6600 2.4GHz (258 Euro)

MB: MSI S775 P35 NEO2-FR 1333MHz 8GB 4xDDRII 1066MHz CF 7.1ch HDA GLAN (112 Euro)
or
Asus S775 P5K Intel P35 DDR2 1066 800 667 1xPCI-E x16 CrossFire X16+X4 SATA II (113 Euro)
or
Asus S775 P5KC Intel P35 DDR2 or DDR3 1xPCI-E SATA II*4 (135 Euro)
HDD: Samsung 160GB 7200rpm SATAII NCQ 8.9ms (54 Euro)

Case+power: Chieftec BH-02B-B-B 400W (94 Euro)
or
Spire SP-3005B Middle ATX Case Black PSU 420W (31 Euro)

Video card: MSI nVidia NX6200TC-TD128E PCI-E 350MHz 500MHz 512MB DDRII 64Bit DirectX 9.0 (34 Euro)(must purchase 'cause the card is not intergarted in mb)

RAM: Kingston 2GB 667MHz DDR2 CL5 DIMM (Kit of 2) (80.5) (How much memory you propose for each machine?)

DVD RW: Samsung DVD-ROM 16x48x beige (1300) ili Pioneer DVDRW DVR112D Black (28 Euro)

all above doubled

Switch: Asus GigaX 1024P 24x10 100; 2x10 100 1000 (109)
or
D-Link DGS-1005D E 5-port Gigabit Unmanaged Switch (31.5 Euro)
or
D-Link DGS-1008D E 8-Port 10 100 1000Mbps Gigabit Unmanaged Switch (46 Euro)


Should I expect aproximately linear speed up with eight cores as I observed on other clusters but with separated single processors? (I saw the pages with jobs run in parallel on core 2 quad but didn't get real answer from the data and those comp. experim.s were runs on one quad core machine: what would be expected with two machines with quad core processor?). What about new Intel P35 chipset: have you pcgamess experience with this new chipset and if so whats the difference between p35 and Blackford one given in performance page (this one is not present on pc market in my country)?

Have a nice day.

Wishes
Zoran Matovic

Message d4763b0bRgK-6473-1308+00.htm, number 141, was posted on Fri Sep 21 '07 at 21:48:25
in reply to 935b012devw-6469-935-00.htm

Re: two quad machines cluster

Alexei Popov
alexei.a.popov@gmail.com


Hi,

1. for two-node clusters, the quality of switch in not very critical. Any gigabit ethernet switch will work (the only exception are very cheap switches).

2. You need to use Gigabit ethernet adapters for your cluster.

3. The quality of motherboard is important as PC GAMESS is a serious stress for system and hardware. The quality of memory is important as well.

4. Try to use a separate HDD for each CPU core.

5. Direct runs will scale very good, and with separate HDDs conventional will do as well.

6. Good cooling system and case are important.

7. Blackford chipset is for servers. It is expensive and requires expensive Xeon processors and FBDIMMs.

Hope this helps.

Alexei Popov




On Mon Sep 17 '07, Zoran Matovic wrote
--------------------------------------
>Wish to ask You: I am planning to build up small cluster of two machines with quadripole processors.
>
>Please have a look the configuration of two identical machines given bellow and help me to decide where different solutions exists:
>
>Processor: Intel LGA Core2Quad Q6600 2.4GHz (258 Euro)
>
>MB: MSI S775 P35 NEO2-FR 1333MHz 8GB 4xDDRII 1066MHz CF 7.1ch HDA GLAN (112 Euro)
>or
>Asus S775 P5K Intel P35 DDR2 1066 800 667 1xPCI-E x16 CrossFire X16+X4 SATA II (113 Euro)
>or
>Asus S775 P5KC Intel P35 DDR2 or DDR3 1xPCI-E SATA II*4 (135 Euro)
>HDD: Samsung 160GB 7200rpm SATAII NCQ 8.9ms (54 Euro)
>
>Case+power: Chieftec BH-02B-B-B 400W (94 Euro)
>or
>Spire SP-3005B Middle ATX Case Black PSU 420W (31 Euro)
>
>Video card: MSI nVidia NX6200TC-TD128E PCI-E 350MHz 500MHz 512MB DDRII 64Bit DirectX 9.0 (34 Euro)(must purchase 'cause the card is not intergarted in mb)
>
>RAM: Kingston 2GB 667MHz DDR2 CL5 DIMM (Kit of 2) (80.5) (How much memory you propose for each machine?)
>
>DVD RW: Samsung DVD-ROM 16x48x beige (1300) ili Pioneer DVDRW DVR112D Black (28 Euro)
>
>all above doubled
>
>Switch: Asus GigaX 1024P 24x10 100; 2x10 100 1000 (109)
>or
>D-Link DGS-1005D E 5-port Gigabit Unmanaged Switch (31.5 Euro)
>or
>D-Link DGS-1008D E 8-Port 10 100 1000Mbps Gigabit Unmanaged Switch (46 Euro)
>
>
>Should I expect aproximately linear speed up with eight cores as I observed on other clusters but with separated single processors? (I saw the pages with jobs run in parallel on core 2 quad but didn't get real answer from the data and those comp. experim.s were runs on one quad core machine: what would be expected with two machines with quad core processor?). What about new Intel P35 chipset: have you pcgamess experience with this new chipset and if so whats the difference between p35 and Blackford one given in performance page (this one is not present on pc market in my country)?
>
>Have a nice day.
>
>Wishes
>Zoran Matovic


Message 51b2da914el-6474-1079-00.htm, number 142, was posted on Sat Sep 22 '07 at 17:59:27
NMR GIAO Method

Adam Sykes
adam.sykes@liv.ac.uk


I can't find any information about PC GAMESS support for NMR calculations based on the GIAO method. This is supported in GAMESS US and I was wondering whether, PC GAMESS can run these calculations or if there was any plans to add this capability in the near future?

Adam Sykes


Message 4df668afbHW-6474-1319+00.htm, number 143, was posted on Sat Sep 22 '07 at 21:59:09
in reply to 51b2da914el-6474-1079-00.htm

Re: NMR GIAO Method

Alex Granovsky
gran@classic.chem.msu.su


Hi,

PC GAMESS does not support NMR calculations at present nor we are planning to implement this feature in the near future. The GAMESS' (US) implementation of GIAO is extremely inefficient so that adding support of NMR calculations to PC GAMESS would require development of the completely new code from the scratch.

Best regards,
Alex Granovsky




On Sat Sep 22 '07, Adam Sykes wrote
-----------------------------------
>I can't find any information about PC GAMESS support for NMR calculations based on the GIAO method. This is supported in GAMESS US and I was wondering whether, PC GAMESS can run these calculations or if there was any plans to add this capability in the near future?

>Adam Sykes


Message 51b2da914el-6475-32+00.htm, number 144, was posted on Sun Sep 23 '07 at 00:31:44
in reply to 4df668afbHW-6474-1319+00.htm

Re^2: NMR GIAO Method

Adam Sykes
adam.sykes@liv.ac.uk


OK. That makes sense. I will have to think again about NMR simulation.

Thank you,

Adam Sykes

On Sat Sep 22 '07, Alex Granovsky wrote
---------------------------------------
>Hi,

>PC GAMESS does not support NMR calculations at present nor we are planning to implement this feature in the near future. The GAMESS' (US) implementation of GIAO is extremely inefficient so that adding support of NMR calculations to PC GAMESS would require development of the completely new code from the scratch.

>Best regards,
>Alex Granovsky
>
>
>
>
>On Sat Sep 22 '07, Adam Sykes wrote
>-----------------------------------
>>I can't find any information about PC GAMESS support for NMR calculations based on the GIAO method. This is supported in GAMESS US and I was wondering whether, PC GAMESS can run these calculations or if there was any plans to add this capability in the near future?

>>Adam Sykes


Message 8afd7fc1aKC-6476-247-00.htm, number 145, was posted on Mon Sep 24 '07 at 04:06:58
DLC and dihedral constraint

Richard
bonarlaw@liv.ac.uk


Is DLC supposed to be compatible with interplane dihedral constraints of the form IFZMAT (1) = 6,I,J,K,L,M (as given in the gamess-us documentation)?

Message d4763b0bbHW-6476-694+00.htm, number 146, was posted on Mon Sep 24 '07 at 11:33:56
in reply to 8afd7fc1aKC-6476-247-00.htm

Re: DLC and dihedral constraint

Alex Granovsky
gran@classic.chem.msu.su


Hi,

no DLCs are not compatible with type=6 coordinates.
They only works with type=1 (bonds), type=2 (angles), and type=3 (torsions) variables.

Regards,
Alex Granovsky


On Mon Sep 24 '07, Richard wrote
--------------------------------
>Is DLC supposed to be compatible with interplane dihedral constraints of the form IFZMAT (1) = 6,I,J,K,L,M (as given in the gamess-us documentation)?
>


Message 558c77a92Sd-6482-1347-00.htm, number 147, was posted on Sun Sep 30 '07 at 22:27:12
functions with angular momenta higher than g

Dmitriy
ltue55@gmail.com


Dear Dr Granovsky,
do you have any plans to implement support for functions with angular momenta higher than g ?

Message c22c517baGH-6483-1046-00.htm, number 148, was posted on Mon Oct 1 '07 at 17:26:29
Memory requiremetns for the new MP2 gradient code

Roman Zubatyuk
zubatyuk@gmail.com


Dear PC GAMESS developers/users.

I got insufficient memory error while performing large MP2 gradient calculation (Nocc=129, Ncore=52, Nvirt=981, distributed over 24 nodes each having 1GB of RAM). There are four options for the code those 'seriously affect amount of memory required': MXI, MXPQ, RSIZE1 and RSIZE2. Regarding this I have two questions:

1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed. Am I right? Should I also change RSIZE1 (or RSIZE2)?

2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

Thanks in advance.
Roman Zubatyuk


Message c22c517baGH-6483-1079-00.htm, number 149, was posted on Mon Oct 1 '07 at 17:58:49
Feature request - meta-GGA DFT functionals

Roman Zubatyuk
zubatyuk@gmail.com


Dear Dr. Granovsky,

Do you have a plans to implement meta-GGA DFT functionals in PC GAMESS? In particular those developed by Truhlar group.

Thanks,
Roman Zubatyuk.


Message d4763b0bbHW-6483-1230+00.htm, number 150, was posted on Mon Oct 1 '07 at 20:29:38
in reply to 558c77a92Sd-6482-1347-00.htm

Re: functions with angular momenta higher than g

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes we plan to incorporate support of h and i harmonics into the next public release of PC GAMESS, version 7.2

Best regards,
Alex Granovsky



On Sun Sep 30 '07, Dmitriy wrote
--------------------------------
>Dear Dr Granovsky,
>do you have any plans to implement support for functions with angular momenta higher than g ?


Message d4763b0bbHW-6483-1243+00.htm, number 151, was posted on Mon Oct 1 '07 at 20:43:02
in reply to c22c517baGH-6483-1046-00.htm

Re: Memory requiremetns for the new MP2 gradient code

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I would recommend to halve RSIZE1 and/or RSIZE2 as the first attempt.


>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed.

This is not exactly the case. The new algorithm is really complex.

MXI affects the amount of disk I/O. Setting MXI to the half of
the number of active occupied orbitals do decrease memory
requirements to some extent but will increase disk I/O.

Smaller MXPQ will increase memory requirements of some
parts of code and decrease requirements of other parts.

>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

The formulas are so complex so that it is much simpler to perform check run. You can use single-node run as an approximate upper limit,
however, parallel run typically uses much less memory. What is even worse, the memory requirements can slightly vary with the molecular geometry.


Best regards,
Alex Granovsky


On Mon Oct 1 '07, Roman Zubatyuk wrote
--------------------------------------
>Dear PC GAMESS developers/users.

>I got insufficient memory error while performing large MP2 gradient calculation (Nocc=129, Ncore=52, Nvirt=981, distributed over 24 nodes each having 1GB of RAM). There are four options for the code those 'seriously affect amount of memory required': MXI, MXPQ, RSIZE1 and RSIZE2. Regarding this I have two questions:

>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed. Am I right? Should I also change RSIZE1 (or RSIZE2)?

>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

>Thanks in advance.
> Roman Zubatyuk


Message d4763b0bbHW-6483-1246+00.htm, number 152, was posted on Mon Oct 1 '07 at 20:46:00
in reply to c22c517baGH-6483-1079-00.htm

Re: Feature request - meta-GGA DFT functionals

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes we plan to implement them in the future but this requires developing of the completely new meta-GGA DFT code framework as the part of PC GAMESS. To add new GGA functional or hybrid GGA functional is easy with the current DFT code, while adding meta-GGA support requires serious additional work.

Regards,
Alex Granovsky


On Mon Oct 1 '07, Roman Zubatyuk wrote
--------------------------------------
>Dear Dr. Granovsky,

> Do you have a plans to implement meta-GGA DFT functionals in PC GAMESS? In particular those developed by Truhlar group.

>Thanks,
> Roman Zubatyuk.


Message c22c517baGH-6485-1281+00.htm, number 153, was posted on Wed Oct 3 '07 at 21:21:29
in reply to d4763b0bbHW-6483-1243+00.htm

Re^2: Memory requiremetns for the new MP2 gradient code

Roman Zubatyuk
zubatyuk@gmail.com


Dear Dr. Granovsky,

I have tryed changing MXPQ and RSIZE2, but it seems that it has no effect on memory needed for second half-transformation. At least as reported by exetyp=check..

all options are default:
# OF WORDS USED FOR MP2 = 123097625
$mp2grd MXPQ=2 rsize2=3000 $end :
# OF WORDS USED FOR MP2 = 123097625

Is it a bug of check procedures?
How can I guess the optimal parameters for large MP2 calculation?

Best regards,
Roman Zubatyuk

On Mon Oct 1 '07, Alex Granovsky wrote
--------------------------------------
>Hi,

>I would recommend to halve RSIZE1 and/or RSIZE2 as the first attempt.
>
>
>>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed.

>This is not exactly the case. The new algorithm is really complex.

>MXI affects the amount of disk I/O. Setting MXI to the half of
>the number of active occupied orbitals do decrease memory
>requirements to some extent but will increase disk I/O.

>Smaller MXPQ will increase memory requirements of some
>parts of code and decrease requirements of other parts.

>>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

>The formulas are so complex so that it is much simpler to perform check run. You can use single-node run as an approximate upper limit,
>however, parallel run typically uses much less memory. What is even worse, the memory requirements can slightly vary with the molecular geometry.
>
>
>Best regards,
>Alex Granovsky
>
>
>On Mon Oct 1 '07, Roman Zubatyuk wrote
>--------------------------------------
>>Dear PC GAMESS developers/users.

>>I got insufficient memory error while performing large MP2 gradient calculation (Nocc=129, Ncore=52, Nvirt=981, distributed over 24 nodes each having 1GB of RAM). There are four options for the code those 'seriously affect amount of memory required': MXI, MXPQ, RSIZE1 and RSIZE2. Regarding this I have two questions:

>>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed. Am I right? Should I also change RSIZE1 (or RSIZE2)?

>>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

>>Thanks in advance.
>> Roman Zubatyuk


Message 4df668afbHW-6485-1334+00.htm, number 154, was posted on Wed Oct 3 '07 at 22:14:15
in reply to c22c517baGH-6485-1281+00.htm

Re^3: Memory requiremetns for the new MP2 gradient code

Alex Granovsky
gran@classic.chem.msu.su


Hi,

rsize1 mainly affects memory needs of the first half-transformation stage.

mxi mainly affects memory needs of the MP2 part of second half-transformation stage.

rsize2 mainly affects memory needs of the DM2 part of second half-transformation stage.

MXPQ cannot be lowered below some threshold defined by the particular basis set in use, and affects DM2 part of the second half-transformation stage as well as 2-e gradient calculation stage.

Turning off all async I/O options will save you some extra memory.

Best regards,
Alex Granovsky


On Wed Oct 3 '07, Roman Zubatyuk wrote
--------------------------------------
>Dear Dr. Granovsky,

> I have tryed changing MXPQ and RSIZE2, but it seems that it has no effect on memory needed for second half-transformation. At least as reported by exetyp=check..

>all options are default:
># OF WORDS USED FOR MP2 = 123097625
>$mp2grd MXPQ=2 rsize2=3000 $end :
># OF WORDS USED FOR MP2 = 123097625

>Is it a bug of check procedures?
>How can I guess the optimal parameters for large MP2 calculation?

>Best regards,
> Roman Zubatyuk
>
>On Mon Oct 1 '07, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>I would recommend to halve RSIZE1 and/or RSIZE2 as the first attempt.
>>
>>
>>>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed.

>>This is not exactly the case. The new algorithm is really complex.

>>MXI affects the amount of disk I/O. Setting MXI to the half of
>>the number of active occupied orbitals do decrease memory
>>requirements to some extent but will increase disk I/O.

>>Smaller MXPQ will increase memory requirements of some
>>parts of code and decrease requirements of other parts.

>>>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

>>The formulas are so complex so that it is much simpler to perform check run. You can use single-node run as an approximate upper limit,
>>however, parallel run typically uses much less memory. What is even worse, the memory requirements can slightly vary with the molecular geometry.
>>
>>
>>Best regards,
>>Alex Granovsky
>>
>>
>>On Mon Oct 1 '07, Roman Zubatyuk wrote
>>--------------------------------------
>>>Dear PC GAMESS developers/users.

>>>I got insufficient memory error while performing large MP2 gradient calculation (Nocc=129, Ncore=52, Nvirt=981, distributed over 24 nodes each having 1GB of RAM). There are four options for the code those 'seriously affect amount of memory required': MXI, MXPQ, RSIZE1 and RSIZE2. Regarding this I have two questions:

>>>1. If I'll halve MXI (or MXPQ), integral half-transformation (first or second) will be performed in two passes and thus approximately twice less memory will be needed. Am I right? Should I also change RSIZE1 (or RSIZE2)?

>>>2. Is there a formula to estimate memory requirement for various steps of MP2 calculation? Or the only solution is to perform check run? Should I run it in parallel, or serial check run will be sufficient?

>>>Thanks in advance.
>>> Roman Zubatyuk


Message 8afd7fc1aKC-6491-119-00.htm, number 155, was posted on Tue Oct 9 '07 at 01:59:47
TDDFT/TDA memory

Richard
bonarlaw@liv.ac.uk


Does anybody have hints for managing memory for largish TDDFT/TDA jobs (3000+ functions with a cc-pVxZ type basis). Seem to need several times the amount of memory recommended by a check run, and even then there may not be enough for fastints/gencon (although some jobs complete despite the memory warnings).

Thanks

Richard


Message d4763b0bbHW-6492-1413+00.htm, number 156, was posted on Wed Oct 10 '07 at 23:33:25
in reply to 8afd7fc1aKC-6491-119-00.htm

Re: TDDFT/TDA memory

Alex Granovsky
gran@classic.chem.msu.su


Hi Richard,

there are two different problems with memory estimate during check runs for TDDFT calculations.

First, fastints/gencon part uses variable amount of memory which cannot be exactly determined without performing actual calculations. Due to this reason fastints/gencon part activates only if there is sufficient memory for it at runtime, otherwise, the old integral code is used. Thus, this should not block calculations entirely but rather can affect their performance.

Second, the memory required for DFT contribution to TDDFT is also the variable part and thus cannot be determined by the check run. At present, check run simply ignores this part of the total memory requirements. The problem is that this contribution is required and thus this variable amount of memory must be available during execution. This is the only point which does can cause problems with memory allocation at runtime. The PC GAMESS v. 7.1.1 will include reasonable estimate of this part of the total required memory during check runs.

Meanwhile, the advises are as follows:
- use as much memory as you can
- if this does not help, try to activate MNMEOP flag in $tddft group


Best regards,
Alex Granovsky




On Tue Oct 9 '07, Richard wrote
-------------------------------
>Does anybody have hints for managing memory for largish TDDFT/TDA jobs (3000+ functions with a cc-pVxZ type basis). Seem to need several times the amount of memory recommended by a check run, and even then there may not be enough for fastints/gencon (although some jobs complete despite the memory warnings).

>Thanks

>Richard


Message 827e0f39JOd-6498-456-00.htm, number 157, was posted on Tue Oct 16 '07 at 07:36:12
Constrained Geometry optimization using cartesian coordinates

mantey
mantey@uiuc.edu


Hello,

Will you be implementing the IFREEZ capability for cartesian coordinates like US Gamess has? (Or is this already possible somehow?) I find this very useful for putting constraints on molecules with large numbers of atoms with symmetry.

Thank you.


Message d4763b0bbHW-6499-1109+00.htm, number 158, was posted on Wed Oct 17 '07 at 18:29:24
in reply to 827e0f39JOd-6498-456-00.htm

Re: Constrained Geometry optimization using cartesian coordinates

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it was already implemented many years ago.

Best regards,
Alex Granovsky


On Tue Oct 16 '07, mantey wrote
-------------------------------
>Hello,

>Will you be implementing the IFREEZ capability for cartesian coordinates like US Gamess has? (Or is this already possible somehow?) I find this very useful for putting constraints on molecules with large numbers of atoms with symmetry.

>Thank you.

>


Message 8afd7fc1aKC-6500-37-00.htm, number 159, was posted on Thu Oct 18 '07 at 00:37:49
DLC again: internal to cartesian convergence warning

Richard
bonarlaw@liv.ac.uk


The following is common for DLC optimizations of our molecules:

WARNING! THE CONVERSION FROM INTERNALS TO CARTESIANS DID NOT CONVERGE IN 10 ITERATIONS

Increasing the number of iterations with ITBMAT mostly eliminates it, but has no effect on time to convergence. Cartesian optimization is still slower than DLC with these warnings. Can they be ignored, or is there something else that could be tweaked to make things better (particularly for large jobs)?

Thanks

Richard


Message 827e0d90JOd-6501-268+00.htm, number 160, was posted on Fri Oct 19 '07 at 04:30:22
in reply to d4763b0bbHW-6499-1109+00.htm

Re^2: Constrained Geometry optimization using cartesian coordinates

mantey
mantey@uiuc.edu


On Wed Oct 17 '07, Alex Granovsky wrote:
>it was already implemented many years ago.


Okay, I finally figured out my mistake. I'll post the information here in case it helps someone else.

As a test case I tried:
$CONTRL UNITS=ANGS COORD=unique RUNTYP=optimize SCFTYP=RHF $END
$STATPT IFREEZ(1)=1 $END
$BASIS GBASIS=N311 NGAUSS=6 NPFUNC=1 NDFUNC=1 $END
$GUESS GUESS=HUCKEL $END
$DATA
Methane
Cnv 2

H 1.0 1.0 0.0000000000 0.8539097649
H 1.0 0.0000000000 1.2077035190 -0.8539097649
C 14.0 0.0000000000 0.0000000000 0.0000000000
$END

Here's a snippet directly from the output:
~~~~snip~~~~
1 ***** EQUILIBRIUM GEOMETRY LOCATED *****
(SUBJECT TO 1 CONSTRAINT(S))

Methane
COORDINATES OF SYMMETRY UNIQUE ATOMS (ANGS)
ATOM CHARGE X Y Z
------------------------------------------------------------
H 1.0 1.2077375961 0.0000000000 0.8539423302
H 1.0 0.0000000000 1.2077852377 -0.8539578896
C 14.0 0.0000000000 0.0000000000 0.0000156567
~~~~snip~~~~

So it recognized that there was a constraint but didn't seem to actually apply it. Looking at the INPUT.DOC it states:

~~~~snip~~~~
IFREEZ = array of internal coordinates to freeze.
For example, IFREEZ(1)=1,3 freezes the two bond
lengths in the $ZMAT example, while optimizing
the angle. You cannot freeze Cartesian coords.
~~~~snip~~~~

Because of that last sentence and the test above I concluded that this was just not possible with the current pcgamess code. As Alex Granovsky mentioned this is not true, pcgamess can do this.

I eventually found out that changing the input text above to use the line:
$STATPT IFREEZ(1)=1,4 $END

It would now work as expected. Despite COORD=unique, it appears you must specify the freeze coordinate # of each coordinate after the symmetry expansion. And it will freeze it to the value assigned during the symmetry expansion.

I hope that helps anyone else in case they run into a similar problem.


Message d4763b0bbHW-6502-1043+00.htm, number 161, was posted on Sat Oct 20 '07 at 17:23:10
in reply to 8afd7fc1aKC-6500-37-00.htm

Re: DLC again: internal to cartesian convergence warning

Alex Granovsky
gran@classic.chem.msu.su


Hi,

One of the possible solutions is to use our gdiis geometry optimizer in conjunction with noreg keyword, e.g.:

$statpt method=gdiis noreg=5 $end

or

$statpt method=gdiis noreg=10 $end

Here, method=gdiis selects gdiis optimizer, while

$statpt noreg= $end

sets the number of geometry optimization cycles between evaluating appropriateness of the DLC regeneration. More precisely, DLCs will always be regenerated if the conversion between internals and Cartesians fails. In addition, every noreg cycles since last DLCs regeneration, the quality of the current DLCs will be examined and corrective actions (DLCs redefinition) will be performed if necessary.

Note that if your molecule undergoes serious geometrical changes during optimization, such as bond breaking, etc..., it is worth to define the problematic bonds in the nonvdw array, as if these bonds disappear at the middle of the run the DLC regeneration algorithm may fail to find the required number of linearly independent variables.


Best regards,
Alex Granovsky


On Thu Oct 18 '07, Richard wrote
--------------------------------
>The following is common for DLC optimizations of our molecules:

>WARNING! THE CONVERSION FROM INTERNALS TO CARTESIANS DID NOT CONVERGE IN 10 ITERATIONS

>Increasing the number of iterations with ITBMAT mostly eliminates it, but has no effect on time to convergence. Cartesian optimization is still slower than DLC with these warnings. Can they be ignored, or is there something else that could be tweaked to make things better (particularly for large jobs)?

>Thanks

>Richard


Message d4763b0bbHW-6502-1096+00.htm, number 161, was edited on Sat Oct 20 '07 at 18:16:37
and replaces message d4763b0bbHW-6502-1043+00.htm

Re: DLC again: internal to cartesian convergence warning

Alex Granovsky
gran@classic.chem.msu.su


Hi,

One of the possible solutions is to use our gdiis geometry optimizer in conjunction with noreg keyword, e.g.:

$statpt method=gdiis noreg=5 $end

or

$statpt method=gdiis noreg=10 $end

Here, method=gdiis selects gdiis optimizer, while

$statpt noreg= < non-negative number > $end

sets the number of geometry optimization cycles between evaluating appropriateness of the DLC regeneration. More precisely, DLCs will always be regenerated if the conversion between internals and Cartesians fails. In addition, every noreg cycles since last DLCs regeneration, the quality of the current DLCs will be examined and corrective actions (DLCs redefinition) will be performed if necessary.

Note that if your molecule undergoes serious geometrical changes during optimization, such as bond breaking, etc..., it is worth to define the problematic bonds in the nonvdw array, as if these bonds disappear at the middle of the run the DLC regeneration algorithm may fail to find the required number of linearly independent variables.


Best regards,
Alex Granovsky


On Thu Oct 18 '07, Richard wrote
--------------------------------
>The following is common for DLC optimizations of our molecules:

>WARNING! THE CONVERSION FROM INTERNALS TO CARTESIANS DID NOT CONVERGE IN 10 ITERATIONS

>Increasing the number of iterations with ITBMAT mostly eliminates it, but has no effect on time to convergence. Cartesian optimization is still slower than DLC with these warnings. Can they be ignored, or is there something else that could be tweaked to make things better (particularly for large jobs)?

>Thanks

>Richard

[ This message was edited on Sat Oct 20 '07 by the author ]


Message c17d4ed5eOC-6519-1232-00.htm, number 162, was posted on Tue Nov 6 '07 at 20:32:52
pcgamess, fc7 and liblam.so.0

Solntsev Pasha
solntsev@univ.kiev.ua


Hi, my question is about library liblam.so.0
I started my job and get message
pcgamess: error while loading shared libraries: libmpi.so.0: cannot open shared object file: No such file or directory

But such files located in nonstandart prefix /lib/oldlib/usrlib64s20
I added this path to LD_LIBRARY_PATH but it is does not wrok property.
how can i solv this problem. Many thanks.


Message d4c0f196bHW-6520-976+00.htm, number 163, was posted on Wed Nov 7 '07 at 16:15:55
in reply to c17d4ed5eOC-6519-1232-00.htm

Re: pcgamess, fc7 and liblam.so.0

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it seems you are trying to use 64-bit libraries with 32-bit PC GAMESS executables. Sure this should not work at all.

Best regards,
Alex Granovsky

On Tue Nov 6 '07, Solntsev Pasha wrote
--------------------------------------
>Hi, my question is about library liblam.so.0
>I started my job and get message
>pcgamess: error while loading shared libraries: libmpi.so.0: cannot open shared object file: No such file or directory

>But such files located in nonstandart prefix /lib/oldlib/usrlib64s20
>I added this path to LD_LIBRARY_PATH but it is does not wrok property.
>how can i solv this problem. Many thanks.

>


Message d92c0be44el-6521-1274-00.htm, number 164, was posted on Thu Nov 8 '07 at 21:14:35
Raman Spectra

Adam Sykes
adam.sykes@liv.ac.uk


Can PC GAMESS produce Raman spectra?


Message 8afd7fc1aKC-6523-104+00.htm, number 165, was posted on Sat Nov 10 '07 at 01:44:10
in reply to d4763b0bbHW-6483-1230+00.htm

Re^2: functions with angular momenta higher than g

Richard
bonarlaw@liv.ac.uk


Will this include support for h terms in ECPs?

Richard

On Mon Oct 1 '07, Alex Granovsky wrote
--------------------------------------
>Hi,

>yes we plan to incorporate support of h and i harmonics into the next public release of PC GAMESS, version 7.2

>Best regards,
>Alex Granovsky
>
>
>
>On Sun Sep 30 '07, Dmitriy wrote
>--------------------------------
>>Dear Dr Granovsky,
>>do you have any plans to implement support for functions with angular momenta higher than g ?


Message 4df668afbHW-6523-1425+00.htm, number 166, was posted on Sat Nov 10 '07 at 23:45:15
in reply to d92c0be44el-6521-1274-00.htm

Re: Raman Spectra

Alex Granovsky
gran@classic.chem.msu.su


Yes it can using runtyp=raman

Regards,
Alex Granovsky

On Thu Nov 8 '07, Adam Sykes wrote
----------------------------------
>Can PC GAMESS produce Raman spectra?

>


Message d4c0f196bHW-6525-950+00.htm, number 167, was posted on Mon Nov 12 '07 at 15:49:59
in reply to 8afd7fc1aKC-6523-104+00.htm

Re^3: functions with angular momenta higher than g

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes they are already supported. Please contact me directly for details.

Best regards,
Alex Granovsky

On Sat Nov 10 '07, Richard wrote
--------------------------------
>Will this include support for h terms in ECPs?

>Richard

>On Mon Oct 1 '07, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>yes we plan to incorporate support of h and i harmonics into the next public release of PC GAMESS, version 7.2

>>Best regards,
>>Alex Granovsky
>>
>>
>>
>>On Sun Sep 30 '07, Dmitriy wrote
>>--------------------------------
>>>Dear Dr Granovsky,
>>>do you have any plans to implement support for functions with angular momenta higher than g ?


Message 864ab85fZ8K-6529-172-00.htm, number 168, was posted on Fri Nov 16 '07 at 02:52:51
Parallel PCGAMESS memory usage

Prasad Yedlapalli
yedlapalli@hotmail.com


Dear PCGamess users,
I am a beginner in using parallel pcgamess. I am running a parallel job on two nodes. Each node has 2GB RAM. However, when I try to run a check job, I get the following error:

ADVANCED RHF-MP2 PROPERTIES/GRADIENT PROGRAM WRITTEN BY ALEX A. GRANOVSKY
------------------------------------ ------------------------------------

# CORE ORBITALS = 43
# OCCUPIED ORBITALS = 177
# BASIS FUNCTIONS = 1423

FIRST MP2 ENERGY/GRADIENT TEI HALF-TRANSFORMATION
# OF WORDS AVAILABLE = 217769836
# OF WORDS USED = 46120036
THRESHOLD FOR KEEPING HALF-TRANSFORMED 2E-INTEGRALS = 1.000E-09


DONE WITH FIRST MP2 GRADIENT TEI HALF-TRANFORMATION

NOT ENOUGH MEMORY FOR SECOND MP2 HALF-TRANSFORMATION
INCREASE MEMORY SIZE BY AT LEAST 58873478 WORDS

Following is my procgrp:
local 0
node23 1 /500GB/pyedlap/pcgathlon/pcgamess

I used the following command to execute the pcgamess:
./pcgamess -f -i 6314h1h_freq.inp -o output.out /500GB/pyedlap/pcgathlon /750GB/pyedlap/pcgathlon -p4pg procgrp &

Following is my input file( not complete):
$contrl scftyp=RHF runtyp=hessian icharg=0 exetyp=check
mplevl=2 mult=1 coord=hint maxit=60 inttyp=hondo icut=11 itol=30 $end
$system timlim=600000 mwords=220 mklnp=2 $end
$basis gbasis=n31 ngauss=6 npfunc=2 diffs=.true. diffsp=.true. $end
$mp2 grdmet=2 method=1 dm2met=3 mnrecl=500 $end
$mp2grd mxpq=2000 rsize2=5000 $end
$p2p p2p=.t. dlb=.t. xdlb=.t. $end
$smp csmtx=.t. $end
$guess guess=huckel $end
$data
6314h1h_freq
C1 0

Even though I am trying to use two nodes with 2GB RAM (total 4GB RAM), PCGAMESS was able to use RAM of one node only. Please let me know how to use memory from both nodes. Is my approach correct?
Many many thanks in advance,
Prasad Yedlapalli


Message 4df668afbHW-6530-1279+00.htm, number 169, was posted on Sat Nov 17 '07 at 21:20:12
in reply to 864ab85fZ8K-6529-172-00.htm

Re: Parallel PCGAMESS memory usage

Alex Granovsky
gran@classic.chem.msu.su


Hi,

> NOT ENOUGH MEMORY FOR SECOND MP2 HALF-TRANSFORMATION
> INCREASE MEMORY SIZE BY AT LEAST 58873478 WORDS

>Even though I am trying to use two nodes with 2GB RAM (total 4GB RAM), PCGAMESS was able to use RAM of one node only. Please let me know how to use memory from both nodes. Is my approach correct?

Actually PC GAMESS uses RAM of both nodes. This memory is not replicated, i.e., it contains different data on different nodes. However, your job is large so that the amount of memory you provided is simply not enough. You can run PC GAMESS on larger number of nodes to reduce memory needs per node. In any case, this is a good idea as your job would require large amount of distributed data on HDDs and on two nodes you may run out of disk space. Next, the calculation time will be much smaller on large number of nodes. Finally, I would suggest you to read the previous discussion on similar topic on these pages (October 01, 2007).

Best regards,
Alex Granovsky




Message c17d4ed5eOC-6533-1215-00.htm, number 170, was posted on Tue Nov 20 '07 at 20:15:40
bsse correction

Solntsev Pasha
solntsev@univ.kiev.ua


Hi
How can i perform dimer computations by eliminating the basis set superposition error using the counterpoise (CP) method of Boys and Bernardi. There are many references about this in the internet, but such procedure does not described step-by-step. Also, i can't found suitable information in the manuals. Many thanks.

Message 8257b3389zt-6534-555-00.htm, number 171, was posted on Wed Nov 21 '07 at 09:15:54
improper scf steps

hohni
ohni@post.kek.jp


Dear PC GAMESS users

I have been executing dft calculation with input files like below.
In the scf procedure, the total energy shift toward improper huge minus value.
I think the reason of this irregular is in reordr=.f. option.
The system has about 3000 AOs and it option is needed to execute the job that has more than about 2000 AOs.
Actually, the job with less than 2000 AOs and reordr=.t., run properly.
Is it possible to eliminate above difficulties?
I appreciate if someone gime me suggestions.

Thank you in advance,

hohni

Input example;
$CONTRL ICHARG=0 MULT=1 SCFTYP=RHF RUNTYP=ENERGY LOCAL=none
COORD=UNIQUE NZVAR=0 MAXIT=50 NPRINT=7 NOSYM=1
PLTORB=.TRUE. EXETYP=RUN IREST=0 dfttyp=svwn1rpa
fstint=.true. reordr=.f. gencon=.true.
$END
$SYSTEM TIMLIM=40000 MEMORY=200000000 kdiag=0 nojac=100 $END
$SCF SHIFT=.TRUE. DAMP=.TRUE. RSTRCT=.FALSE.
DIIS=.true. soscf=.false. DIRSCF=.TRUE. NCONV=3 ETHRSH=2.0 $END
$BASIS GBASIS=STO NGAUSS=3 $END
$GUESS GUESS=HUCKEL $END
$INTGRL QFMM=.false. SCHWRZ=.TRUE. $END
$FMM METHOD=FULLDRCT LEX=0 $END
$DATA
******
C1 0




Message 4df668afbHW-6534-813+00.htm, number 172, was posted on Wed Nov 21 '07 at 13:32:47
in reply to 8257b3389zt-6534-555-00.htm

Re: improper scf steps

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please send me your compressed exact input file and we'll try to reproduce your problem and find the solution for you.

Best regards,
Alex Granovsky



On Wed Nov 21 '07, hohni wrote
------------------------------
>Dear PC GAMESS users

>I have been executing dft calculation with input files like below.
>In the scf procedure, the total energy shift toward improper huge minus value.
>I think the reason of this irregular is in reordr=.f. option.
>The system has about 3000 AOs and it option is needed to execute the job that has more than about 2000 AOs.
>Actually, the job with less than 2000 AOs and reordr=.t., run properly.
>Is it possible to eliminate above difficulties?
>I appreciate if someone gime me suggestions.

>Thank you in advance,

>hohni

>Input example;
> $CONTRL ICHARG=0 MULT=1 SCFTYP=RHF RUNTYP=ENERGY LOCAL=none
> COORD=UNIQUE NZVAR=0 MAXIT=50 NPRINT=7 NOSYM=1
> PLTORB=.TRUE. EXETYP=RUN IREST=0 dfttyp=svwn1rpa
> fstint=.true. reordr=.f. gencon=.true.
> $END
> $SYSTEM TIMLIM=40000 MEMORY=200000000 kdiag=0 nojac=100 $END
> $SCF SHIFT=.TRUE. DAMP=.TRUE. RSTRCT=.FALSE.
> DIIS=.true. soscf=.false. DIRSCF=.TRUE. NCONV=3 ETHRSH=2.0 $END
> $BASIS GBASIS=STO NGAUSS=3 $END
> $GUESS GUESS=HUCKEL $END
> $INTGRL QFMM=.false. SCHWRZ=.TRUE. $END
> $FMM METHOD=FULLDRCT LEX=0 $END
> $DATA
> ******
> C1 0
>
>
>
>


Message 58d654d7eOC-6535-0-00.htm, number 173, was posted on Thu Nov 22 '07 at 00:01:08
bsse correction 2

Solntsev Pasha
solntsev@univ.kiev.ua


Hi.
I study week interaction between two molecules. Geometry optimization and full energy minimization ware performed and interaction energy was computed by formula delatE = E(AB)-E(A)-E(B). But it is very importent to include BSSE correction. The RYNTYP=MOROKUMA does not produce an accurate meaning of the interaction energy which adopt very small value (from morokuma decomposition) (I checked this on well known example from literature). Also, i can't use BSSE correction from Morokuma decomposition it self, because it is not a full Boys counterpoise correction and lead to very small value of E(BSSE) (compared with literature). Best solution for such kind of problem is to make Counterpoise Method (CP). But how can i make it in PC GAMESS? Many thanks.


Message c317208ebQn-6535-1015+00.htm, number 174, was posted on Thu Nov 22 '07 at 16:55:08
in reply to 58d654d7eOC-6535-0-00.htm

Re: bsse correction 2

Pedro Silva
pedros@ufp.pt


On Thu Nov 22 '07, Solntsev Pasha wrote
---------------------------------------
Dear Pasha,

I have placed a demonstration for He2 on my homepage. You can download it from:

www2.ufp.pt/~pedros/science/bsse.zip

Basically, you should prepare three files. The first one (let's call it "complete molecule") has all of your atoms and basis sets. The others have all basis sets but only a set of the atoms (let's call them "fragment A" and "fragment B"). These latter files should be prepared in ther following way:
a) prepare your files as usual, using COORD=UNIQUE. Then
b) The basis set should be included explicitly, with all exponents, etc. in $DATA, rather than in $BASIS. Basis sets can be conveniently downloaded from http://www.emsl.pnl.gov/forms/basisform.html
c) Substitute the charge of the atoms which should be absent from the calculation but whose basis functions should be kept.

Run your three files. The real interaction energy between the fragments A and B should be :

FINAL ENERGY "complete molecule" - FINAL ENERGY "fragment A" -FINAL ENERGY "fragment B".

If you compute the energies for separated fragments A and B WITHOUT the extra basis functions on the absent atoms, you get the wrong interaction energy. The difference between the two computed interaction energies is the BSSE.

I hope I have been clear (and that there are no major errors in this post :-)

Pedro Silva

http://www2.ufp.pt/~pedros/science/science.htm


Message 58d67755eOC-6536-102+00.htm, number 175, was posted on Fri Nov 23 '07 at 01:42:17
in reply to c317208ebQn-6535-1015+00.htm

Re^2: bsse correction 2

Solntsev Pasha
solntsev@univ.kiev.ua


Dear Pedro, many, many thanks. It is very good tutorial and i have solved my problem.



Message 4df668afbHW-6536-1019+00.htm, number 176, was posted on Fri Nov 23 '07 at 17:02:07
in reply to c317208ebQn-6535-1015+00.htm

Re^2: bsse correction 2

Alex Granovsky
gran@classic.chem.msu.su


Dear Pedro,

thanks for answering Pasha's question. I'm attaching the archive with your sample to this message and hope it will be useful for others.

I would like to mention the alternative way - you can just specify the negative nuclear charge to ignore that nuclea in calculations but put proper basis functions on its place automatically.

Best regards,
Alex


On Thu Nov 22 '07, Pedro Silva wrote
------------------------------------
>On Thu Nov 22 '07, Solntsev Pasha wrote
>---------------------------------------
>Dear Pasha,

>I have placed a demonstration for He2 on my homepage. You can download it from:

> www2.ufp.pt/~pedros/science/bsse.zip

>Basically, you should prepare three files. The first one (let's call it "complete molecule") has all of your atoms and basis sets. The others have all basis sets but only a set of the atoms (let's call them "fragment A" and "fragment B"). These latter files should be prepared in ther following way:
>a) prepare your files as usual, using COORD=UNIQUE. Then

>b) The basis set should be included explicitly, with all exponents, etc. in $DATA, rather than in $BASIS. Basis sets can be conveniently downloaded from

http://www.emsl.pnl.gov/forms/basisform.html

c) Substitute the charge of the atoms which should be absent from the calculation but whose basis functions should be kept.

>Run your three files. The real interaction energy between the fragments A and B should be :

>FINAL ENERGY "complete molecule" - FINAL ENERGY "fragment A" -FINAL ENERGY "fragment B".

>If you compute the energies for separated fragments A and B WITHOUT the extra basis functions on the absent atoms, you get the wrong interaction energy. The difference between the two computed interaction energies is the BSSE.

>I hope I have been clear (and that there are no major errors in this post :-)

>Pedro Silva

http://www2.ufp.pt/~pedros/science/science.htm

This message contains the 17 kb attachment
[ bsse.zip ] BSSE calculation sample


Message 864ab85fZ8K-6548-1214-00.htm, number 177, was posted on Wed Dec 5 '07 at 20:13:51
Deletion of stored intermediate files: MOINTS

Prasad Yedlapalli
yedlapalli@hotmail.com


Dear PCGAMESS users,
I was running an hessian job with the following command:
./pcgamess -f -i 3214h1h_freq.inp -o 3214h1h_freq1205.out -ex /750GB/pyedlap/pcgathlon &

It created some 50 intermediate files named MOINTS. Afterwards, I killed the job using kill -9 command. I changed some input options and I restarted the job with the following command:
./pcgamess -f -i 3214h1h_freq.inp -o 3214h1h_freq1205.out -ex /750GB/pyedlap/pcgathlon &

Then I killed the job and there was only one MOINTS file. I just wanted to make sure enough hard drive space was there. I used the df command to verify storage space. Following is the partial output:
/dev/hdc1 732549636 103808692 628740944 15% /750GB (1K-blocks)
The size of MOINTS is only 1343312.
-rw-rw-r-- 1 pyedlap pyedlap 1343312 Dec 5 10:13 MOINTS

Now I have the following situation:
I don't see any of the old MOINTS files ( 50 in number). They just disappeared. I used ls -a command. Still, my hard drive is 15% full. It used to be 1% full. Would you please suggest how to remove the unseen MOINTS files to recover the hard drive space?

Thank you very much for your help,
Prasad


Message 864ab85fZ8K-6548-1396+00.htm, number 178, was posted on Wed Dec 5 '07 at 23:16:06
in reply to 864ab85fZ8K-6548-1214-00.htm

Re: Deletion of stored intermediate files: MOINTS

Prasad Yedlapalli
yedlapalli@hotmail.com


Dear PCGAMESS users,
Actually there were pcgamess processes running in the background, which I didn't know. Once I killed all those processes, the disk space became available.
Thank you,
Prasad


On Wed Dec 5 '07, Prasad Yedlapalli wrote
-----------------------------------------
>Dear PCGAMESS users,
> I was running an hessian job with the following command:
> ./pcgamess -f -i 3214h1h_freq.inp -o 3214h1h_freq1205.out -ex /750GB/pyedlap/pcgathlon &

>It created some 50 intermediate files named MOINTS. Afterwards, I killed the job using kill -9 command. I changed some input options and I restarted the job with the following command:
> ./pcgamess -f -i 3214h1h_freq.inp -o 3214h1h_freq1205.out -ex /750GB/pyedlap/pcgathlon &

> Then I killed the job and there was only one MOINTS file. I just wanted to make sure enough hard drive space was there. I used the df command to verify storage space. Following is the partial output:
> /dev/hdc1 732549636 103808692 628740944 15% /750GB (1K-blocks)
> The size of MOINTS is only 1343312.
> -rw-rw-r-- 1 pyedlap pyedlap 1343312 Dec 5 10:13 MOINTS
>
> Now I have the following situation:
> I don't see any of the old MOINTS files ( 50 in number). They just disappeared. I used ls -a command. Still, my hard drive is 15% full. It used to be 1% full. Would you please suggest how to remove the unseen MOINTS files to recover the hard drive space?
>
> Thank you very much for your help,
> Prasad


Message 935b012dcvZ-6553-828-00.htm, number 179, was posted on Mon Dec 10 '07 at 13:49:08
running jobs on 12cpus demands more time than 8 cpus?

Zoran Matovic
zmatovic@kg.ac.yu


Dear pcgamess users,
Till couple days ago I ran pcgamess on 8cpu cluster. now I extended it for 4cpu more. However, when ran jobs with 12cpus it costs me two times more time than with 8 cpu. Why is that so? Can anyone help?
Thanks in advance
zoran matovic

cluster: three machines connected through gigabite switch.
each machine:
Processor: CPU INTEL CORE 2 QUAD Q6600 2.4GHz
MB: Gigabyte GA-P35-DS3P
RAM: MEMORIJA DDR2 transcdent 1024MB 800MHz 4GB
HDD 160 GB Western Digital WD1600AAJS S-ATA II 2
Video: VGA MSI PCI-E NX7300LE-TD256E 2
Optics: DVD RW PIONEER DVR-112DBk 2
Case: CHIEFTEC 550W


OS: Scientific Linux 4.5 PBS Torgue (use pcgamess 7.01)


Message 4df668afbHW-6553-1100+00.htm, number 180, was posted on Mon Dec 10 '07 at 18:20:09
in reply to 935b012dcvZ-6553-828-00.htm

Re: running jobs on 12cpus demands more time than 8 cpus?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

possibly you are bandwithd limited.

Are you using ch_p4 device and running PC GAMESS in DLB mode over P2P interface?

Best regards,
Alex Granovsky


On Mon Dec 10 '07, Zoran Matovic wrote
--------------------------------------
>Dear pcgamess users,
>Till couple days ago I ran pcgamess on 8cpu cluster. now I extended it for 4cpu more. However, when ran jobs with 12cpus it costs me two times more time than with 8 cpu. Why is that so? Can anyone help?
>Thanks in advance
>zoran matovic

>cluster: three machines connected through gigabite switch.
>each machine:
>Processor: CPU INTEL CORE 2 QUAD Q6600 2.4GHz
>MB: Gigabyte GA-P35-DS3P
>RAM: MEMORIJA DDR2 transcdent 1024MB 800MHz 4GB
>HDD 160 GB Western Digital WD1600AAJS S-ATA II 2
>Video: VGA MSI PCI-E NX7300LE-TD256E 2
>Optics: DVD RW PIONEER DVR-112DBk 2
>Case: CHIEFTEC 550W
>
>
>OS: Scientific Linux 4.5 PBS Torgue (use pcgamess 7.01)
>


Message 4df668afbHW-6553-1172+00.htm, number 180, was edited on Mon Dec 10 '07 at 19:32:27
and replaces message 4df668afbHW-6553-1100+00.htm

Re: running jobs on 12cpus demands more time than 8 cpus?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

possibly you are bandwidth limited, if your job is not large enough.

Are you using ch_p4 device and running PC GAMESS in DLB mode over P2P interface?

Best regards,
Alex Granovsky


On Mon Dec 10 '07, Zoran Matovic wrote
--------------------------------------
>Dear pcgamess users,
>Till couple days ago I ran pcgamess on 8cpu cluster. now I extended it for 4cpu more. However, when ran jobs with 12cpus it costs me two times more time than with 8 cpu. Why is that so? Can anyone help?
>Thanks in advance
>zoran matovic

>cluster: three machines connected through gigabite switch.
>each machine:
>Processor: CPU INTEL CORE 2 QUAD Q6600 2.4GHz
>MB: Gigabyte GA-P35-DS3P
>RAM: MEMORIJA DDR2 transcdent 1024MB 800MHz 4GB
>HDD 160 GB Western Digital WD1600AAJS S-ATA II 2
>Video: VGA MSI PCI-E NX7300LE-TD256E 2
>Optics: DVD RW PIONEER DVR-112DBk 2
>Case: CHIEFTEC 550W
>
>
>OS: Scientific Linux 4.5 PBS Torgue (use pcgamess 7.01)
>

[ This message was edited on Mon Dec 10 '07 by the author ]


Message 935b012bcvZ-6553-1190+00.htm, number 181, was posted on Mon Dec 10 '07 at 19:51:21
in reply to 4df668afbHW-6553-1172+00.htm

Re^2: running jobs on 12cpus demands more time than 8 cpus?

Zoran Matovic
zmatovic@kg.ac.yu


Thanks for answer Alex. I must admitt that I'm not some expert for linux and writter of scripts; however my colleagues wrote a script for me to be able to start pcgamess in parallel using PBS torgue as a job manager. the scripts is given bellow:
#!/bin/sh

### Set the job name
#PBS -N batch

### Declare myprogram non-rerunable
#PBS -r n

### Have PBS mail you results
##PBS -m ae

#PBS -q batch

### Specify the number of cpus for your job. This example will run on 16 cpus
### using 8 nodes with 2 processes per node. Set this to a max of 2 and 2 if you
### are using the "bigmem" queue.
##PBS -l nodes=1:ppn=4
#PBS -l nodes=3:ppn=4

INPUT="newpdda"
TPMDIRLIST=""
echo Working directory is $PBS_O_WORKDIR
cd $PBS_O_WORKDIR
echo Running on host `hostname`
echo Time is `date`
echo Directory is `pwd`
echo This jobs runs on the following processors:
echo `cat $PBS_NODEFILE`
NPROCS=`wc -l < $PBS_NODEFILE`
echo This job has allocated $NPROCS cpus

NODES=`cat $PBS_NODEFILE`

for host in ${NODES}
do
N=$RANDOM
echo "Copying input file to $host..."
ssh -v $host rm -rf /tmp/pcgamess_$N/*
ssh -v $host mkdir /tmp/pcgamess_$N
ssh -v $host cp $PWD/$INPUT /tmp/pcgamess_$N/input
TMPDIRLIST=$TMPDIRLIST' '/tmp/pcgamess_$N
done

echo $TMPDIRLIST

mpirun -np $NPROCS -machinefile $PBS_NODEFILE pcgamess $TMPDIRLIST > $INPUT.log

It is not quite clear for me where to put -ex /home/zoran/pcgamess (in my case; is it written OK?). Will you help me how to handle this to activate p2p. When I activate p2p and wrote -ex home at the and of above scripts the pcgamess abort with a message: Loading P2P interface library... error loading library!

Thanks.

zoran



On Mon Dec 10 '07, Alex Granovsky wrote
---------------------------------------
>Hi,

>possibly you are bandwidth limited, if your job is not large enough.

>Are you using ch_p4 device and running PC GAMESS in DLB mode over P2P interface?

>Best regards,
>Alex Granovsky
>
>
>On Mon Dec 10 '07, Zoran Matovic wrote
>--------------------------------------
>>Dear pcgamess users,
>>Till couple days ago I ran pcgamess on 8cpu cluster. now I extended it for 4cpu more. However, when ran jobs with 12cpus it costs me two times more time than with 8 cpu. Why is that so? Can anyone help?
>>Thanks in advance
>>zoran matovic

>>cluster: three machines connected through gigabite switch.
>>each machine:
>>Processor: CPU INTEL CORE 2 QUAD Q6600 2.4GHz
>>MB: Gigabyte GA-P35-DS3P
>>RAM: MEMORIJA DDR2 transcdent 1024MB 800MHz 4GB
>>HDD 160 GB Western Digital WD1600AAJS S-ATA II 2
>>Video: VGA MSI PCI-E NX7300LE-TD256E 2
>>Optics: DVD RW PIONEER DVR-112DBk 2
>>Case: CHIEFTEC 550W
>>
>>
>>OS: Scientific Linux 4.5 PBS Torgue (use pcgamess 7.01)
>>


Message c06b52111oc-6554-1083-00.htm, number 182, was posted on Tue Dec 11 '07 at 18:02:47
Question about CEP-31G* basis set

Pasquale
morvillo@portici.enea.it


Hi,
I started to use PCGAMESS some days ago.
I want to make a DFT calculation using 6-31G* and CEP-31G* basis set (I found an article whare the author use Gaussian with these basis set), but I got results very far from the results reported inside that article when I use SBKJC with GAMESS.
If I use 6-31G* basis set, everything is as expected.
I thought that CEP-31G (Gaussian) corresponds to SBKJC (GAMESS); is it true?
My input file is (thiophene molecule):

$CONTRL RUNTYP=ENERGY SCFTYP=RHF DFTTYP=B3LYP ECP=SBKJC COORD=CART $END
$BASIS GBASIS=SBKJC NDFUNC=1 $END

$DATA
C4H4S
C1
C 6.0 -0.081780 -1.233739 0.000000
C 6.0 -1.324229 -0.718399 0.000000
C 6.0 -1.324314 0.718246 0.000000
C 6.0 -0.081929 1.233751 0.000000
S 16.0 1.124115 0.000079 0.000000
H 1.0 0.199785 -2.267156 0.000000
H 1.0 -2.216044 -1.316435 0.000000
H 1.0 -2.216212 1.316178 0.000000
H 1.0 0.199503 2.267192 0.000000
$END

Can you give me some hints?

Best regards.


Message 4df668afbHW-6555-897+00.htm, number 183, was posted on Wed Dec 12 '07 at 14:56:50
in reply to 935b012bcvZ-6553-1190+00.htm

Re^3: running jobs on 12cpus demands more time than 8 cpus?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please ask your friend to add the following command line switch to the pc gamess command line options:

pcgamess ...some other switches... -ex < path to directory where pcgp2p.ex resides >

E.g., if you have PC GAMESS' runtime extensions (pcgp2p.ex and fastdiag.ex) in the /opt/pcgamess71 folder on all nodes, use the following switch:

 -ex /opt/pcgamess71

Then, add the following line to your input:

 $p2p p2p=.t. dlb=.t. $end

Best regards,
Alex


On Mon Dec 10 '07, Zoran Matovic wrote
--------------------------------------
>Thanks for answer Alex. I must admitt that I'm not some expert for linux and writter of scripts; however my colleagues wrote a script for me to be able to start pcgamess in parallel using PBS torgue as a job manager. the scripts is given bellow:
>#!/bin/sh

>### Set the job name
>#PBS -N batch

>### Declare myprogram non-rerunable
>#PBS -r n

>### Have PBS mail you results
>##PBS -m ae

>#PBS -q batch

>### Specify the number of cpus for your job. This example will run on 16 cpus
>### using 8 nodes with 2 processes per node. Set this to a max of 2 and 2 if you
>### are using the "bigmem" queue.
>##PBS -l nodes=1:ppn=4
>#PBS -l nodes=3:ppn=4

>INPUT="newpdda"
>TPMDIRLIST=""
>echo Working directory is $PBS_O_WORKDIR
>cd $PBS_O_WORKDIR
>echo Running on host `hostname`
>echo Time is `date`
>echo Directory is `pwd`
>echo This jobs runs on the following processors:
>echo `cat $PBS_NODEFILE`
>NPROCS=`wc -l < $PBS_NODEFILE`
>echo This job has allocated $NPROCS cpus

>NODES=`cat $PBS_NODEFILE`

>for host in ${NODES}
>do
> N=$RANDOM
> echo "Copying input file to $host..."
> ssh -v $host rm -rf /tmp/pcgamess_$N/*
> ssh -v $host mkdir /tmp/pcgamess_$N
> ssh -v $host cp $PWD/$INPUT /tmp/pcgamess_$N/input
> TMPDIRLIST=$TMPDIRLIST' '/tmp/pcgamess_$N
>done

>echo $TMPDIRLIST

>mpirun -np $NPROCS -machinefile $PBS_NODEFILE pcgamess $TMPDIRLIST > $INPUT.log
>

>It is not quite clear for me where to put -ex /home/zoran/pcgamess (in my case; is it written OK?). Will you help me how to handle this to activate p2p. When I activate p2p and wrote -ex home at the and of above scripts the pcgamess abort with a message: Loading P2P interface library... error loading library!

>Thanks.

>zoran
>
>
>
>On Mon Dec 10 '07, Alex Granovsky wrote
>---------------------------------------
>>Hi,

>>possibly you are bandwidth limited, if your job is not large enough.

>>Are you using ch_p4 device and running PC GAMESS in DLB mode over P2P interface?

>>Best regards,
>>Alex Granovsky
>>
>>
>>On Mon Dec 10 '07, Zoran Matovic wrote
>>--------------------------------------
>>>Dear pcgamess users,
>>>Till couple days ago I ran pcgamess on 8cpu cluster. now I extended it for 4cpu more. However, when ran jobs with 12cpus it costs me two times more time than with 8 cpu. Why is that so? Can anyone help?
>>>Thanks in advance
>>>zoran matovic

>>>cluster: three machines connected through gigabite switch.
>>>each machine:
>>>Processor: CPU INTEL CORE 2 QUAD Q6600 2.4GHz
>>>MB: Gigabyte GA-P35-DS3P
>>>RAM: MEMORIJA DDR2 transcdent 1024MB 800MHz 4GB
>>>HDD 160 GB Western Digital WD1600AAJS S-ATA II 2
>>>Video: VGA MSI PCI-E NX7300LE-TD256E 2
>>>Optics: DVD RW PIONEER DVR-112DBk 2
>>>Case: CHIEFTEC 550W
>>>
>>>
>>>OS: Scientific Linux 4.5 PBS Torgue (use pcgamess 7.01)
>>>


Message c06b52111oc-6556-822+00.htm, number 184, was posted on Thu Dec 13 '07 at 13:41:40
in reply to c06b52111oc-6554-1083-00.htm

Re: Question about CEP-31G* basis set

Pasquale
morvillo@portici.enea.it


Hi,
I found the problem with the GAMESS equivalent of Gaussian ECP-31G* basis set.
It was not a problem of GAMESS calculation, but a problem of Mask program I am using to see output file!
The ECP removes the core electrons (and same number of protons), but Mask doesn't take in acount this case!
So I toke the data of my interest directly from output file and the results are near what reported in literature using Gaussian!

Message 5901ca31d5e-6558-174-00.htm, number 185, was posted on Sat Dec 15 '07 at 02:54:14
Facio documentation

Amit
amit1b@hotmail.com


Hi,
I couldn't find any documentation (manual, tutorial, etc.) on Facio's website. Would you happen to know where I can find it?
Thanks,
Amit.

Message 5901ca31d5e-6558-1127-00.htm, number 186, was posted on Sat Dec 15 '07 at 18:49:04
GAMESS optimization error

Amit
amit1b@hotmail.com


I have been using GAMESS (via Facio) to optimize a small pdb file containing the coordinates of NADPH (enzyme cofactor, see attached file). GAMESS starts running, but then the aborts with the following error message:

Execution of GAMESS terminated abnormally
*** PC GAMESS Output***
*** Failure to Locate Stationary Point ***

Can anybody explain this?
Thanks,
Amit.

This message contains the 6 kb attachment
[ NADPH2.pdb ] PDB-style coordinates of NADPH, prepared by GAMESS before optimization


Message 589dda9abQn-6559-73+00.htm, number 187, was posted on Sun Dec 16 '07 at 01:13:02
in reply to 5901ca31d5e-6558-1127-00.htm

Re: GAMESS optimization error

Pedro Silva
pedros@ufp.pt


On Sat Dec 15 '07, Amit wrote
-----------------------------
>I have been using GAMESS (via Facio) to optimize a small pdb file containing the coordinates of NADPH (enzyme cofactor, see attached file). GAMESS starts running, but then the aborts with the following error message:

>Execution of GAMESS terminated abnormally
>*** PC GAMESS Output***
>*** Failure to Locate Stationary Point ***

>Can anybody explain this?
>Thanks,
>Amit.


You must include the following line in your gamess input file:

$statpt nstep=xxx $end

(with xxx substituted by the number of optimixation steps you plan to use. I usually set it to 150, and use the last set of coordinates as input to a new job, in case it has not converged)


Pedro Silva

PS: Gamess syntax is actually very easy to learn. I learnt Gamess by studying a few of the test jobs, and checking the manual for each command.The manual is very clearly written, and I find the refs.doc file very helpful in understanding the underlying theory. I found that a graphical user interace (GUI) is only needed to build your molecule, and by writing the Gamess comands ourselves we can precisely tune it to do what we want, not simply what the helpful people who wrote the GUI thought we might want....


Message 5901ca31d5e-6559-103+00.htm, number 188, was posted on Sun Dec 16 '07 at 01:45:03
in reply to 589dda9abQn-6559-73+00.htm

Re^2: GAMESS optimization error

Amit
amit1b@hotmail.com


Hi Pedro,
Thanks for your help. Actually, I have this line in the .inp file (see line 7 in attachment). Could you please take a quick look and see if anything is wrong with the file?
Best,
Amit.
PC. I agree with you about not being dependent on the GUI. I used it mainly to turn my pdb-style coordinates to GAMESS compatible file.


On Sun Dec 16 '07, Pedro Silva wrote
------------------------------------
>On Sat Dec 15 '07, Amit wrote
>-----------------------------
>>I have been using GAMESS (via Facio) to optimize a small pdb file containing the coordinates of NADPH (enzyme cofactor, see attached file). GAMESS starts running, but then the aborts with the following error message:

>>Execution of GAMESS terminated abnormally
>>*** PC GAMESS Output***
>>*** Failure to Locate Stationary Point ***

>>Can anybody explain this?
>>Thanks,
>>Amit.
>
>
>You must include the following line in your gamess input file:

> $statpt nstep=xxx $end

>(with xxx substituted by the number of optimixation steps you plan to use. I usually set it to 150, and use the last set of coordinates as input to a new job, in case it has not converged)
>
>
>Pedro Silva

>PS: Gamess syntax is actually very easy to learn. I learnt Gamess by studying a few of the test jobs, and checking the manual for each command.The manual is very clearly written, and I find the refs.doc file very helpful in understanding the underlying theory. I found that a graphical user interace (GUI) is only needed to build your molecule, and by writing the Gamess comands ourselves we can precisely tune it to do what we want, not simply what the helpful people who wrote the GUI thought we might want....

This message contains the 4 kb attachment
[ NADPH.inp ] Input file for GAMESS, prepared by Facio.


Message 4df668afbHW-6559-814+00.htm, number 189, was posted on Sun Dec 16 '07 at 13:33:55
in reply to 5901ca31d5e-6559-103+00.htm

Re^3: GAMESS optimization error

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it seems your structure has several hydrogen atoms missing.
I would also recommend you to use DIIS or SOSCF SCF convergers.

Regards,
Alex Granovsky



On Sun Dec 16 '07, Amit wrote
-----------------------------
>Hi Pedro,
>Thanks for your help. Actually, I have this line in the .inp file (see line 7 in attachment). Could you please take a quick look and see if anything is wrong with the file?
>Best,
>Amit.
>PC. I agree with you about not being dependent on the GUI. I used it mainly to turn my pdb-style coordinates to GAMESS compatible file.
>
>
>On Sun Dec 16 '07, Pedro Silva wrote
>------------------------------------
>>On Sat Dec 15 '07, Amit wrote
>>-----------------------------
>>>I have been using GAMESS (via Facio) to optimize a small pdb file containing the coordinates of NADPH (enzyme cofactor, see attached file). GAMESS starts running, but then the aborts with the following error message:

>>>Execution of GAMESS terminated abnormally
>>>*** PC GAMESS Output***
>>>*** Failure to Locate Stationary Point ***

>>>Can anybody explain this?
>>>Thanks,
>>>Amit.
>>
>>
>>You must include the following line in your gamess input file:

>> $statpt nstep=xxx $end

>>(with xxx substituted by the number of optimixation steps you plan to use. I usually set it to 150, and use the last set of coordinates as input to a new job, in case it has not converged)
>>
>>
>>Pedro Silva

>>PS: Gamess syntax is actually very easy to learn. I learnt Gamess by studying a few of the test jobs, and checking the manual for each command.The manual is very clearly written, and I find the refs.doc file very helpful in understanding the underlying theory. I found that a graphical user interace (GUI) is only needed to build your molecule, and by writing the Gamess comands ourselves we can precisely tune it to do what we want, not simply what the helpful people who wrote the GUI thought we might want....


Message 589dda9abQn-6559-897+00.htm, number 190, was posted on Sun Dec 16 '07 at 14:56:51
in reply to 5901ca31d5e-6559-103+00.htm

Re^3: GAMESS optimization error

Pedro Silva
pedros@ufp.pt


NADPH has -4 charge, but your input file has ICHARG=0. I have tried to converge it with ICHARG=-4, but it did not work. Neither did it work with 4 extra hydrogens attached to the phosphates and ICHARG=0.
I guess you might try to split you NAPDH molecule in two according to the process described in the tutorial

http://classic.chem.msu.su/gran/gamess/

Pedro S


Message 4df668afbHW-6559-1037+00.htm, number 191, was posted on Sun Dec 16 '07 at 17:17:50
in reply to 589dda9abQn-6559-897+00.htm

Re^4: GAMESS optimization error

Alex Granovsky
gran@classic.chem.msu.su


Hi Amit and Pedro,

I was able to converge it by adding four hydrogens by the (almost) arbitrary manner, changing hamiltonian to AM1 and forcing DIIS.

The input file I used is attached.

Best regards,
Alex




On Sun Dec 16 '07, Pedro Silva wrote
------------------------------------
>NADPH has -4 charge, but your input file has ICHARG=0. I have tried to converge it with ICHARG=-4, but it did not work. Neither did it work with 4 extra hydrogens attached to the phosphates and ICHARG=0.
>I guess you might try to split you NAPDH molecule in two according to the process described in the tutorial

>http://classic.chem.msu.su/gran/gamess/http://classic.chem.msu.su/gran/gamess/

>Pedro S

This message contains the 7 kb attachment
[ input ] Modified input file


Message 5901ca31d5e-6559-1045+00.htm, number 192, was posted on Sun Dec 16 '07 at 17:25:14
in reply to 4df668afbHW-6559-1037+00.htm

Re^5: GAMESS optimization error

Amit
amit1b@hotmail.com


Thanks! I'm just wondering about the extra hydrogens, since the phosphates should be deprotonated at physiological pH.
Best,
Amit.


On Sun Dec 16 '07, Alex Granovsky wrote
---------------------------------------
>Hi Amit and Pedro,

>I was able to converge it by adding four hydrogens by the (almost) arbitrary manner, changing hamiltonian to AM1 and forcing DIIS.

>The input file I used is attached.

>Best regards,
>Alex
>
>
>
>
>On Sun Dec 16 '07, Pedro Silva wrote
>------------------------------------
>>NADPH has -4 charge, but your input file has ICHARG=0. I have tried to converge it with ICHARG=-4, but it did not work. Neither did it work with 4 extra hydrogens attached to the phosphates and ICHARG=0.
>>I guess you might try to split you NAPDH molecule in two according to the process described in the tutorial

>>http://classic.chem.msu.su/gran/gamess/http://classic.chem.msu.su/gran/gamess/http://classic.chem.msu.su/gran/gamess/http://classic.chem.msu.su/gran/gamess/

>>Pedro S


Message 589dda9abQn-6559-1279+00.htm, number 193, was posted on Sun Dec 16 '07 at 21:19:32
in reply to 5901ca31d5e-6559-1045+00.htm

Re^6: GAMESS optimization error

Pedro Silva
pedros@ufp.pt


I am attaching an optimized structure of NADPH, with correct H (I hope..)

Pedro S

This message contains the 6 kb attachment
[ NADPH_1.inp ]


Message 5901ca31d5e-6559-1340+00.htm, number 194, was posted on Sun Dec 16 '07 at 22:20:35
in reply to 589dda9abQn-6559-1279+00.htm

Re^7: GAMESS optimization error

Amit
amit1b@hotmail.com


Thanks Pedro!

On Sun Dec 16 '07, Pedro Silva wrote
------------------------------------
>I am attaching an optimized structure of NADPH, with correct H (I hope..)

>Pedro S


Message 558cee36O52-6563-761-00.htm, number 195, was posted on Thu Dec 20 '07 at 12:41:38
Functional with asymptotic behavior correction

Khimich
khimich@kinet.chem.msu.ru


There is functional with asymptotic behavior correction in PC GAMESS?

Message 4df668afbHW-6563-952+00.htm, number 196, was posted on Thu Dec 20 '07 at 15:52:03
in reply to 558cee36O52-6563-761-00.htm

Re: Functional with asymptotic behavior correction

Alex Granovsky
gran@classic.chem.msu.su


Hi,

No there is no asymptotically corrected functionals support at present in the PC GAMESS.
However, we are working on the implementation of self-interaction correction as the part of PC GAMESS' DFT code. Corrected functionals will automatically have correct long-range behavior.

Best regards,
Alex Granovsky


On Thu Dec 20 '07, Khimich wrote
--------------------------------
>There is functional with asymptotic behavior correction in PC GAMESS?


Message 935b012dcvZ-6564-770-00.htm, number 197, was posted on Fri Dec 21 '07 at 12:52:02
Problems with new 7.1 version in beowulf cluster jobs

Zoran Matovic
zmatovic@kg.ac.yu


Dear Pcgamess users,
Just two days ago I replaced old 7.01 pcgamess on my 12 cpu beowulf cluster (three intel core 2 quad; in total 12 cpu; pbs torque batch system; 12 GB ram in total for three machines) with new 7.1 version. However, when started job with selfconstructed script (see attachment) pcgamess run with n-1 processors instead of n requested. Note that with the same script old 7.01 version works perfectly.

Thanking in advance I wish all of you marry christmass and happy new 2008. year.

zoran

This message contains the 2 kb attachment
[ p.sub ] submission script


Message 4df668afbHW-6566-911-00.htm, number 198, was posted on Sun Dec 23 '07 at 15:11:14
PC GAMESS version 7.1 to version 7.1.5 update information

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users!

The update to PC GAMESS version 7.1.5 is available.

Please visit the "Download" section of the PC GAMESS homepage for more information and for download links.

We recommend all PC GAMESS users to upgrade to version 7.1.5.

Best regards,
Alex Granovsky


Message 4df668afbHW-6566-911+00.htm, number 198, was edited on Sun Dec 23 '07 at 20:59:44
and replaces message 4df668afbHW-6566-911-00.htm

PC GAMESS version 7.1 to version 7.1.5 update information

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users!

The update to PC GAMESS version 7.1.5 is now available for download.

Please visit the "Downloads" section of the PC GAMESS homepage for more information and for download links.

We recommend all PC GAMESS users to upgrade to version 7.1.5.

Best regards,
Alex Granovsky

[ This message was edited on Sun Dec 23 '07 by the author ]


Message 7c6678a9ZGa-6567-948-00.htm, number 199, was posted on Mon Dec 24 '07 at 15:48:30
EXECUTION OF GAMESS TERMINATED ABNORMALLY

venkataramanan
venkataramananns@gmail.com


I am trying to optimize a 50 atom containing structure. While execution i am getting the following error

EXECUTION OF GAMESS TERMINATED ABNORMALLY AT XXXX at XXXX
Optimization failed! Too many steps taken
Restart with last coordinates and increase
NSTEP in group $STATPT


Even after increasing the NSTEP from 5000 to 500000 i am getting the same error. The optimization has not crossed the first step.


I would be thankful if some one exaplin how to receify the problem

With thanks in advance

This message contains the 243 kb attachment
[ gamess.out ] OUT put


Message 4df668afbHW-6567-969+00.htm, number 200, was posted on Mon Dec 24 '07 at 16:08:58
in reply to 7c6678a9ZGa-6567-948-00.htm

Re: EXECUTION OF GAMESS TERMINATED ABNORMALLY

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the corrected input for your job is as follows:

! Minimize (Energy/Geometry) B3LYP/DZV     
 $CONTRL ICHARG=2 MAXIT=100 MULT=3 RUNTYP=OPTIMIZE 
 SCFTYP=UHF DFTTYP=B3LYP5 $END           
 $SYSTEM MWORDS=20 $END           
 $BASIS GBASIS=TZV POLAR=POPLE $END           
 $SCF DIRSCF=.true. $END           
 $STATPT nstep=1000 $end
 $DATA          
enc             
C1            
.....

Best regards,
Alex Granovsky


Message 4df668afbHW-6567-1051+00.htm, number 200, was edited on Mon Dec 24 '07 at 17:31:05
and replaces message 4df668afbHW-6567-969+00.htm

Re: EXECUTION OF GAMESS TERMINATED ABNORMALLY

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the corrected input for your job is as follows:

! Minimize (Energy/Geometry) B3LYP/DZV     
 $CONTRL ICHARG=2 MAXIT=100 MULT=3 RUNTYP=OPTIMIZE 
 SCFTYP=UHF DFTTYP=B3LYP5 $END           
 $SYSTEM MWORDS=20 TIMLIM=10000 $END           
 $BASIS GBASIS=TZV POLAR=POPLE $END           
 $SCF DIRSCF=.true. $END           
 $STATPT nstep=1000 $end
 $DATA          
enc             
C1            
.....

Best regards,
Alex Granovsky

[ This message was edited on Mon Dec 24 '07 by the author ]


Message 7c6678a9ZGa-6568-116+00.htm, number 201, was posted on Tue Dec 25 '07 at 01:57:00
in reply to 4df668afbHW-6567-1051+00.htm

Re^2: EXECUTION OF GAMESS TERMINATED ABNORMALLY

venkataramanan
venkataramananns@gmail.com


Dear Dr, Granovsky,

Thank you very much for your reply. I corrected the input file as directed by you and run the calcuation. I once again got error due to Memory problem. Then in gave the comand
$SYSTEM
MWORDS=12
$END
and rerun the calcuation. Once again i got the same memeory error. I am enclosing both the input and output files for you suggestion.

WIth thanks in advance

WIsh you a very happy X-Mass

Sincerely
Venkataramanan


On Mon Dec 24 '07, Alex Granovsky wrote
---------------------------------------
>Hi,

>the corrected input for your job is as follows:

>

! Minimize (Energy/Geometry) B3LYP/DZV     
 $CONTRL ICHARG=2 MAXIT=100 MULT=3 RUNTYP=OPTIMIZE 
 SCFTYP=UHF DFTTYP=B3LYP5 $END           
 $SYSTEM MWORDS=20 TIMLIM=10000 $END           
 $BASIS GBASIS=TZV POLAR=POPLE $END           
 $SCF DIRSCF=.true. $END           
 $STATPT nstep=1000 $end
 $DATA          
enc             
C1            
.....

>Best regards,
>Alex Granovsky
>

This message contains the 47 kb attachment
[ Input and Output.rar ] RAR file, Please kindly unzip


Message d42eff1ebHW-6568-992+00.htm, number 202, was posted on Tue Dec 25 '07 at 16:33:11
in reply to 7c6678a9ZGa-6568-116+00.htm

Re^3: EXECUTION OF GAMESS TERMINATED ABNORMALLY

Alex Granovsky
gran@classic.chem.msu.su


Hi,

attached please find corrected input file.

Best regards,
Alex Granovsky

This message contains the 2 kb attachment
[ corrected_input.rar ] Input file with mistakes corrected


Message 4df668afbHW-6569-1188+00.htm, number 203, was posted on Wed Dec 26 '07 at 19:48:19
in reply to 935b012dcvZ-6564-770-00.htm

Re: Problems with new 7.1 version in beowulf cluster jobs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this seems to be strange to me.

Please verify that you can run PC GAMESS on all cpus manually,
using the correctly written procgrp file, e.g.:

local 3
host2... 4 /home/zoran/pcgamess710/pcgamess
host3... 4 /home/zoran/pcgamess710/pcgamess

and providing -p4pg mpich command line option;
and that PC GAMESS behaves as expected.

Next, I would highly recommend to use new syntax and the -ex
command line option instead of creating/passing list of working
directories manually.


Best regards,
Alex Granovsky


Message 4df668afbHW-6569-1312+00.htm, number 203, was edited on Wed Dec 26 '07 at 21:52:14
and replaces message 4df668afbHW-6569-1188+00.htm

Re: Problems with new 7.1 version in beowulf cluster jobs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this looks strange to me.

Please verify that you can run PC GAMESS on all cpus manually,
using the correctly written procgrp file, e.g.:

local 3
host2... 4 /home/zoran/pcgamess710/pcgamess
host3... 4 /home/zoran/pcgamess710/pcgamess

and providing -p4pg mpich command line option;
and that PC GAMESS behaves as expected.

Next, I would highly recommend to use new syntax and the -ex
command line option instead of creating/passing list of working
directories manually.


Best regards,
Alex Granovsky

[ This message was edited on Wed Dec 26 '07 by the author ]


Message a12df4d7Z8K-6572-1421-00.htm, number 204, was posted on Sat Dec 29 '07 at 23:41:27
PCGAMESS under OpenSUSE Linux

Anatoliy Volkov
avolkov@mtsu.edu


Greetings,

I can't seem to be able to run the latest version of PCGAMESS under OpenSUSE Linux. I've downloaded archive
pcgamess_linux_mpich_p4.zip and after unzipping it I've got 'pcg710.exe' file. I then changed attributes of this file as 'chmod a+x pcg710.exe'. However, this exe file does not produce any output.

For example, I use the following GAMESS input file called 'test.inp':

$contrl scftyp=RHF runtyp=ENERGY mplevl=2 units=ANGS icharg=0 mult=1 $end
$basis gbasis=N31 ngauss=6 ndfunc=1 nffunc=0 npfunc=1 diffsp=.TRUE. $end
$guess guess=huckel $end
$data
Water MP2/6-31+G(d,p) single point energy calculation
CNv 2

O 8.0 0.00000000 0.00000000 0.11761300
H 1.0 0.00000000 0.75734800 -0.47045000
$end

I run PCGMESS as follows:

/tmp/gamess> ls
pcg710.exe* test.inp

/tmp/gamess> time ./pcg710.exe -i test.inp -o test.out
0.000u 0.000s 0:00.00 0.0% 0+0k 0+0io 0pf+0w

/tmp/gamess> ls
pcg710.exe* test.inp

It seems that PCGAMESS does perform any calculation at all, even though no error is printed out. It does not matter what input file I use or how I name it - no output is ever printed and no calculation is performed.

PCGAMESS exe file seems to be compatible with my OS:

/tmp/gamess> file ./pcg710.exe
./pcg710.exe: ELF 32-bit LSB executable, Intel 80386, version 1, statically linked, stripped

My operating system is Linux OpenSUSE 10.3. I've tried running PCGAMESS on two different systems:

1) dual-cpu Core2Duo running 32-bit OpenSUSE 10.3

/tmp/gamess> uname -a
Linux galaxy 2.6.22.13-0.3-default #1 SMP 2007/11/19 15:02:58 UTC i686 i686 i386 GNU/Linux

/tmp/gamess> cat /etc/SuSE-release
openSUSE 10.3 (i586)
VERSION = 10.3

2) dual-socket quad-core Xeon (8 cpus total) running 64-bit OpenSUSE 10.3.

/tmp/gamess> uname -a
Linux voltron 2.6.22.13-0.3-default #1 SMP 2007/11/19 15:02:58 UTC x86_64 x86_64 x86_64 GNU/Linux

/tmp/gamess> cat /etc/SuSE-release
openSUSE 10.3 (X86-64)
VERSION = 10.3

but get the same problem on both systems.

I also tried other PCGAMESS linux exe files (statically & dynamically linked, and for different cpus) but I get the same problem.

Am I doing something wrong here ?

Thank you in advance for your help!

Best Regards,
Anatoliy


Message d42eebbabHW-6573-783+00.htm, number 205, was posted on Sun Dec 30 '07 at 13:03:32
in reply to a12df4d7Z8K-6572-1421-00.htm

Re: PCGAMESS under OpenSUSE Linux

Alex Granovsky
gran@classic.chem.msu.su


Dear Anatoliy,

The pcg710.exe file is the rar self-extracting archive. We supposed that the version of self-extractor we used is compatible with any version of Linux, it seems it really is not. Please, try to unrar pcg710.exe manually and follow the displayed instructions. When prompted, enter the second password. Several files will be extracted, including file called pcgamess - the main PC GAMESS executable. Do not forget to issue "chmod a+x pcgamess".

Best regards,
Alex Granovsky









On Sat Dec 29 '07, Anatoliy Volkov wrote
----------------------------------------
>Greetings,

>I can't seem to be able to run the latest version of PCGAMESS under OpenSUSE Linux. I've downloaded archive
>pcgamess_linux_mpich_p4.zip and after unzipping it I've got 'pcg710.exe' file. I then changed attributes of this file as 'chmod a+x pcg710.exe'. However, this exe file does not produce any output.

>For example, I use the following GAMESS input file called 'test.inp':

>$contrl scftyp=RHF runtyp=ENERGY mplevl=2 units=ANGS icharg=0 mult=1 $end
>$basis gbasis=N31 ngauss=6 ndfunc=1 nffunc=0 npfunc=1 diffsp=.TRUE. $end
>$guess guess=huckel $end
>$data
>Water MP2/6-31+G(d,p) single point energy calculation
>CNv 2

>O 8.0 0.00000000 0.00000000 0.11761300
>H 1.0 0.00000000 0.75734800 -0.47045000
>$end

>I run PCGMESS as follows:

>/tmp/gamess> ls
>pcg710.exe* test.inp

>/tmp/gamess> time ./pcg710.exe -i test.inp -o test.out
>0.000u 0.000s 0:00.00 0.0% 0+0k 0+0io 0pf+0w

>/tmp/gamess> ls
>pcg710.exe* test.inp

>It seems that PCGAMESS does perform any calculation at all, even though no error is printed out. It does not matter what input file I use or how I name it - no output is ever printed and no calculation is performed.

>PCGAMESS exe file seems to be compatible with my OS:

>/tmp/gamess> file ./pcg710.exe
>./pcg710.exe: ELF 32-bit LSB executable, Intel 80386, version 1, statically linked, stripped

>My operating system is Linux OpenSUSE 10.3. I've tried running PCGAMESS on two different systems:

>1) dual-cpu Core2Duo running 32-bit OpenSUSE 10.3

>/tmp/gamess> uname -a
>Linux galaxy 2.6.22.13-0.3-default #1 SMP 2007/11/19 15:02:58 UTC i686 i686 i386 GNU/Linux

>/tmp/gamess> cat /etc/SuSE-release
>openSUSE 10.3 (i586)
>VERSION = 10.3

>2) dual-socket quad-core Xeon (8 cpus total) running 64-bit OpenSUSE 10.3.

>/tmp/gamess> uname -a
>Linux voltron 2.6.22.13-0.3-default #1 SMP 2007/11/19 15:02:58 UTC x86_64 x86_64 x86_64 GNU/Linux

>/tmp/gamess> cat /etc/SuSE-release
>openSUSE 10.3 (X86-64)
>VERSION = 10.3

>but get the same problem on both systems.

>I also tried other PCGAMESS linux exe files (statically & dynamically linked, and for different cpus) but I get the same problem.

>Am I doing something wrong here ?

>Thank you in advance for your help!

>Best Regards,
>Anatoliy


Message 574ed00dUaS-6574-658-00.htm, number 206, was posted on Mon Dec 31 '07 at 10:58:45
help with parallel benchmark

volker
v.karuth@netcologne.de


Hi Alex,

# i recently upgraded my personal system to a Q6600 Kentsfield processor and Vista-OS. Could you provide me the input files of your Kentsfield-Performance tests, as you did with the PC GAMESS v. 7.0 benchmarks on different Intel Pentium 4 family (Netburst) processors.

I doubt if my calculation is running only on 1 node not 4 if compairing to your timings. Just to find out what i am doing wrong, i attached my input-file.

Thanks a lot

Volker from Cologne, Gemany
-----------------
Answer from Alex Granovsky:

Hi,

you should use mklnp=4 only for tests #2 and #4 as they are the only running in multithreaded mode
as explained on the page you mentioned. For other tests, you need to:

1. Backup you present mpibind.dll file found in the PC GAMESS' home directory.
2. Replace it by the MPIBIND.NT-MPICH-SMP.DLL file found in the BINDINGS folder.
Note you must manually rename it to mpibind.dll
3. Check if the mpich_smp.dll file is in the PC GAMESS' home directory (it should be there).
4. Read documentation on how to run PC GAMESS in parallel, including at least the following:

readme.commandline
readme.parallel.windows.wmpi
and finally,
readme.parallel.windows.nt-mpich-smp

5. Run PC GAMESS in parallel as described in readme.parallel.windows.nt-mpich-smp using proper command line switches. For four cores, use -np 4 at the end of command line options.

Best regards,
Alex Granovsky
P.S. Please repost your question to the PC GAMESS' Discussion pages at MSU - I'll repost my answer here too, as it can be helpful for other people too. Please let me know on any possible problems with your registration on the forum.

This message contains the 95 kb attachment
[ Q6850.out ] output_file


Message 574fd022UaS-6574-748+00.htm, number 207, was posted on Mon Dec 31 '07 at 12:28:09
in reply to 574ed00dUaS-6574-658-00.htm

Re: help with parallel benchmark

volker
v.karuth@netcologne.de


Thanks for the fast advice !

Now, gamess starts 4 times, and everything seems fine, my commandline is
"pcgamess.exe -f >C:\Users\Volker\Q6850_mklnp4_2.out C:\volker\dir1 C:\volker\dir2 C:\volker\dir3 C:\volker\dir4 -np 4"
previously i created the four working directories, but gamess ended either with

"DIRECT SCF CALCULATION, SCHWRZ=T FDIFF=T
FSF: fatal error no. 0x00000020 in sub SEQOPN on unit 90

EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 10:08:30 LT 31-DE"

or going into nirvana during the direct scf-calculation
(this means job is no more running, working-directories are empty, but outputfile stopps at the start of direct-scf, with no error)

I doubt if i had to start a service manually under VISTA ? How cani find out if mpich or wmpi is needed/installed and running?
But the readme-file readme.parallel.windows.nt-mpich-smp said that no addittional packages are needed ?

I stepped further but i need another advice, anyone using intel-quadcore and vista out-there, any comments on the commandline switches ?

Best regards
Volker


On Mon Dec 31 '07, volker wrote
-------------------------------
>Hi Alex,

># i recently upgraded my personal system to a Q6600 Kentsfield processor and Vista-OS. Could you provide me the input files of your Kentsfield-Performance tests, as you did with the PC GAMESS v. 7.0 benchmarks on different Intel Pentium 4 family (Netburst) processors.

>I doubt if my calculation is running only on 1 node not 4 if compairing to your timings. Just to find out what i am doing wrong, i attached my input-file.

> Thanks a lot

>Volker from Cologne, Gemany
>-----------------
>Answer from Alex Granovsky:

>Hi,
>
>you should use mklnp=4 only for tests #2 and #4 as they are the only running in multithreaded mode
>as explained on the page you mentioned. For other tests, you need to:
>
>1. Backup you present mpibind.dll file found in the PC GAMESS' home directory.
>2. Replace it by the MPIBIND.NT-MPICH-SMP.DLL file found in the BINDINGS folder.
> Note you must manually rename it to mpibind.dll
>3. Check if the mpich_smp.dll file is in the PC GAMESS' home directory (it should be there).
>4. Read documentation on how to run PC GAMESS in parallel, including at least the following:
>
>readme.commandline
>readme.parallel.windows.wmpi
> and finally,
>readme.parallel.windows.nt-mpich-smp
>
>5. Run PC GAMESS in parallel as described in readme.parallel.windows.nt-mpich-smp using proper command line switches. For four cores, use -np 4 at the end of command line options.
>
>Best regards,
>Alex Granovsky
>P.S. Please repost your question to the PC GAMESS' Discussion pages at MSU - I'll repost my answer here too, as it can be helpful for other people too. Please let me know on any possible problems with your registration on the forum.
>
>


Message 4df668afbHW-6574-1014+00.htm, number 208, was posted on Mon Dec 31 '07 at 16:54:17
in reply to 574fd022UaS-6574-748+00.htm

Re^2: help with parallel benchmark

Alex Granovsky
gran@classic.chem.msu.su


Hi,

if you want to use the old-style syntax, the proper command line is as follows:

pcgamess.exe -f C:\volker\dir1 C:\volker\dir2 C:\volker\dir3 C:\volker\dir4 -np 4 > C:\Users\Volker\Q6850_mklnp4_2.out

i.e., the I/O redirection should be the last in the command line.

Otherwise, your input redirection as parsed by cmd.exe terminates the list of command line options to be passed to PC GAMESS. This results in wrong directories list passed to PC GAMESS and file i/o conflicts between different PC GAMESS processes.

Alternatively, you can use, e.g.:

pcgamess.exe -f -o C:\Users\Volker\Q6850_mklnp4_2.out C:\volker\dir1 C:\volker\dir2 C:\volker\dir3 C:\volker\dir4 -np 4

Or, using the new syntax:

pcgamess.exe -r -f -p -i C:\Users\Volker\Q6850_mklnp4_2.inp -o C:\Users\Volker\Q6850_mklnp4_2.out -t C:\volker\tmp -np 4


Best regards,
Alex Granovsky






On Mon Dec 31 '07, volker wrote
-------------------------------
>Thanks for the fast advice !

>Now, gamess starts 4 times, and everything seems fine, my commandline is
>"pcgamess.exe -f >C:\Users\Volker\Q6850_mklnp4_2.out C:\volker\dir1 C:\volker\dir2 C:\volker\dir3 C:\volker\dir4 -np 4"
>previously i created the four working directories, but gamess ended either with

>"DIRECT SCF CALCULATION, SCHWRZ=T FDIFF=T
> FSF: fatal error no. 0x00000020 in sub SEQOPN on unit 90

> EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 10:08:30 LT 31-DE"

>or going into nirvana during the direct scf-calculation
>(this means job is no more running, working-directories are empty, but outputfile stopps at the start of direct-scf, with no error)

>I doubt if i had to start a service manually under VISTA ? How can i find out if mpich or wmpi is needed/installed and running?
>But the readme-file readme.parallel.windows.nt-mpich-smp said that no addittional packages are needed ?

>I stepped further but i need another advice, anyone using intel-quadcore and vista out-there, any comments on the commandline switches ?

>Best regards
>Volker


Message 584aaa057ZI-6579-159-00.htm, number 209, was posted on Sat Jan 5 '08 at 02:39:13
JOB HAS EXHAUSTED ITS CPU ALLOTMENT

pheidrias
pcgamess@pheidrias.info


Hello everybody!

Does anybody know, what this message means?

I'm running pcgamess (Serial/parallel Linux binaries) on the command-line using single processor on a dual-core AMD-System with Ubuntu 7.10.

after

./pcgamess -r -f -p -i inputfile.inp -o outputfile.log &


I get:

...... END OF 2-ELECTRON GRADIENT ......



CPU TIME: STEP = 398.95 , TOTAL = 36146.6 SECONDS ( 602.4 MIN)

WALL CLOCK TIME: STEP = 399.00 , TOTAL = 54996.6 SECONDS ( 916.6 MIN)

CPU UTILIZATION: STEP = 99.99%, TOTAL = 65.73%

**** JOB HAS EXHAUSTED ITS CPU ALLOTMENT ****

36000. -5 20 9 0 0 1 4 1 1 1 1 186933 1 4998


in the optimization procedure.

How can I avoid this?
PCGamess should only be able to use one of my two processors - so there couldn't be any overrun?

Thanks for help,
pheidrias


Message 4df668afbHW-6579-984+00.htm, number 210, was posted on Sat Jan 5 '08 at 16:23:55
in reply to 584aaa057ZI-6579-159-00.htm

Re: JOB HAS EXHAUSTED ITS CPU ALLOTMENT

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you need to increase the value of timlim variable in the $system group. The default value is 600, i.e., 600 minutes or 10 hours. For large runs, you can use e.g., the following:


 $system timlim=10000 $end

Best regards,
Alex Granovsky


Message 4df668afbHW-6579-986+00.htm, number 211, was posted on Sat Jan 5 '08 at 16:26:41
in reply to 9e68a878O5K-6270-1397-00.htm

Re: Upgrade to NBO 5.x

Alex Granovsky
gran@classic.chem.msu.su


Hi all,

finally we have the beta version to test with NBO 5.0 incorporated into it.

Best regards,
Alex



On Fri Mar 2 '07, James A Hoobler wrote
---------------------------------------
>Hi Alex,

>Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

>Thanks.

>James

>


Message 4df668afbHW-6579-988+00.htm, number 211, was edited on Sat Jan 5 '08 at 16:28:08
and replaces message 4df668afbHW-6579-986+00.htm

Re: Upgrade to NBO 5.x

Alex Granovsky
gran@classic.chem.msu.su


Hi all,

finally we have the beta version of PC GAMESS to test with NBO 5.0 incorporated into it.

Best regards,
Alex



On Fri Mar 2 '07, James A Hoobler wrote
---------------------------------------
>Hi Alex,

>Sorry to press you on this issue, but when do you plan to upgrade the NBO code from 4.M to 5.x?

>Thanks.

>James

>

[ This message was edited on Sat Jan 5 '08 by the author ]


Message a4293c46KOU-6585-875-00.htm, number 212, was posted on Fri Jan 11 '08 at 14:35:42
Overlap Matrix

Heibbe
heibbe@ueg.br


PC GAMESS users,


I source a keyword that printed the overlap matrix. However I did not
find. It is possible for the PCGAMESS users to get the overlap matrix?

Best regards

Heibbe


Message a5f69a21RsX-6586-381-00.htm, number 213, was posted on Sat Jan 12 '08 at 06:21:29
error with remote command

ahmed
ahmed912005@yahoo.com


PC GAMESS users,
i can not run parallel job on two nodes(windows) , i get these error messages
error with remote command
can not connect to chem002:10060 (chem002 is the slave node hostname)
p0_2096: net_creat_slave by remote shell:0
here is what i did
1- i installed PCGAMESS and the included WMPI-1.3 on both nodes
2- i did a successful serial job on both of them separately
3- i checked the connection between the two nodes by ping "hostname"
4- i made sure that mpi_service.exe is running in system services in both nodes

i would be thankful if someone can tell me why i can not run a parallel job ?

thanks in advance
Best Regards


Message 3b9725e639w-6586-431-00.htm, number 214, was posted on Sat Jan 12 '08 at 07:11:27
Spin-orbit coupling(SOC)

Yubing.Si
yubing.si@gmail.com


Hi all,

I am trying to calculate the spin-orbit coupling constants(SOCC). While execution i am getting the following error:
*** CHECK YOUR INPUT CHARGE AND MULTIPLICITY ***
THERE ARE ?204 ELECTRONS, WITH CHARGE ICHARG= 0
BUT YOU SELECTED MULTIPLICITY MULT= 6

Even after change the MULTIPLICITY from 6 to 5 (though is incorrect in the system of bare mental of Nb and CH4) i am getting the new error:

*** ERROR ***
AN UNEXPECTED $END WAS ENCOUNTERED IN $ECP GROUP.
EVERY ATOM IN THE MOLECULE MUST BE DEFINED IN $ECP.

Why does these happen?

Many thanks.

Yubing.Si


Message 4df668afbHW-6586-1249+00.htm, number 215, was posted on Sat Jan 12 '08 at 20:48:54
in reply to 3b9725e639w-6586-431-00.htm

Re: Spin-orbit coupling(SOC)

Alexei Popov
alexei.a.popov@gmail.com


Hi,

please attach exact input file to your post.


Best regards,
Alex Popov


On Sat Jan 12 '08, Yubing.Si wrote
----------------------------------
>Hi all,
>
>I am trying to calculate the spin-orbit coupling constants(SOCC). While execution i am getting the following error:
> *** CHECK YOUR INPUT CHARGE AND MULTIPLICITY ***
> THERE ARE ?204 ELECTRONS, WITH CHARGE ICHARG= 0
> BUT YOU SELECTED MULTIPLICITY MULT= 6

>Even after change the MULTIPLICITY from 6 to 5 (though is incorrect in the system of bare mental of Nb and CH4) i am getting the new error:

> *** ERROR ***
> AN UNEXPECTED $END WAS ENCOUNTERED IN $ECP GROUP.
> EVERY ATOM IN THE MOLECULE MUST BE DEFINED IN $ECP.

>Why does these happen?

>Many thanks.
>
>Yubing.Si
>


Message 4df668afbHW-6586-1251+00.htm, number 216, was posted on Sat Jan 12 '08 at 20:51:23
in reply to a5f69a21RsX-6586-381-00.htm

Re: error with remote command

Alex Granovsky
gran@classic.chem.msu.su


Hi,

check if there is a firewall blocking PC GAMESS.

Note that the recommended MPI implementation for PC GAMESS is NT-MPICH, not WMPI 1.3

Best regards,
Alex Granovsky



On Sat Jan 12 '08, ahmed wrote
------------------------------
>PC GAMESS users,
>i can not run parallel job on two nodes(windows) , i get these error messages
>error with remote command
>can not connect to chem002:10060 (chem002 is the slave node hostname)
>p0_2096: net_creat_slave by remote shell:0
>here is what i did
>1- i installed PCGAMESS and the included WMPI-1.3 on both nodes
>2- i did a successful serial job on both of them separately
>3- i checked the connection between the two nodes by ping "hostname"
>4- i made sure that mpi_service.exe is running in system services in both nodes

>i would be thankful if someone can tell me why i can not run a parallel job ?

>thanks in advance
>Best Regards
>


Message 3b9725e639w-6587-269+00.htm, number 217, was posted on Sun Jan 13 '08 at 04:34:43
in reply to 4df668afbHW-6586-1249+00.htm

Re^2: Spin-orbit coupling(SOC)

Yubing.Si
yubing.si@gmail.com


On Sat Jan 12 '08, Alexei Popov wrote
-------------------------------------
>Hi,

>please attach exact input file to your post.
>
>
>Best regards,
>Alex Popov
>
>
>On Sat Jan 12 '08, Yubing.Si wrote
>----------------------------------
>>Hi all,
>>
>>I am trying to calculate the spin-orbit coupling constants(SOCC). While execution i am getting the following error:
>> *** CHECK YOUR INPUT CHARGE AND MULTIPLICITY ***
>> THERE ARE ?204 ELECTRONS, WITH CHARGE ICHARG= 0
>> BUT YOU SELECTED MULTIPLICITY MULT= 6

>>Even after change the MULTIPLICITY from 6 to 5 (though is incorrect in the system of bare mental of Nb and CH4) i am getting the new error:

>> *** ERROR ***
>> AN UNEXPECTED $END WAS ENCOUNTERED IN $ECP GROUP.
>> EVERY ATOM IN THE MOLECULE MUST BE DEFINED IN $ECP.

>>Why does these happen?

>>Many thanks.
>>
>>Yubing.Si
>>

This message contains the 54 kb attachment
[ nb-3.inp ] This is the input file,with the Basis set($ECP) is Stuttgart RSC 1997 ECP v1.1.


Message 4df668afbHW-6587-935+00.htm, number 218, was posted on Sun Jan 13 '08 at 15:35:55
in reply to 3b9725e639w-6587-269+00.htm

Re^3: Spin-orbit coupling(SOC)

Alexei Popov
alexei.a.popov@gmail.com


Hi,

your input contains multiple mistakes.

I'm attaching the sample input file (which is not exactly what you wanted to do) just for your reference.

Regards,
Alex Popov


On Sun Jan 13 '08, Yubing.Si wrote
----------------------------------
>On Sat Jan 12 '08, Alexei Popov wrote
>-------------------------------------
>>Hi,

>>please attach exact input file to your post.
>>
>>
>>Best regards,
>>Alex Popov
>>
>>
>>On Sat Jan 12 '08, Yubing.Si wrote
>>----------------------------------
>>>Hi all,
>>>
>>>I am trying to calculate the spin-orbit coupling constants(SOCC). While execution i am getting the following error:
>>> *** CHECK YOUR INPUT CHARGE AND MULTIPLICITY ***
>>> THERE ARE ?204 ELECTRONS, WITH CHARGE ICHARG= 0
>>> BUT YOU SELECTED MULTIPLICITY MULT= 6

>>>Even after change the MULTIPLICITY from 6 to 5 (though is incorrect in the system of bare mental of Nb and CH4) i am getting the new error:

>>> *** ERROR ***
>>> AN UNEXPECTED $END WAS ENCOUNTERED IN $ECP GROUP.
>>> EVERY ATOM IN THE MOLECULE MUST BE DEFINED IN $ECP.

>>>Why does these happen?

>>>Many thanks.
>>>
>>>Yubing.Si
>>>

This message contains the 54 kb attachment
[ soc_sample_incorrect.inp ] just a sample file for SOC calculations


Message c1e90450Lgz-6587-1137-00.htm, number 219, was posted on Sun Jan 13 '08 at 18:57:10
Mixed cluster/SMP execution

Ivan Fedyanin
octy@xrlab.ineos.ac.ru


Hi all,
I'm trying to execute pcgamess on a dual-core AMD64-based linux cluster.

First of all, I use the following version: [version 7.1 (Tornado), build number 4444].
Yeah, I do know that this is a beta-release, but it's the one I can find to work with
lam-mpi 7.x.x.

DFT-code works pretty well if only one core of the processor is to be used. But as long
as I put corresponding instruction (cpu=2) in my file, describing the nodes, program
prints the following error:

FSF: fatal error no. 0x04000000 in sub PACKREAD on unit 8

My guess is that this error is concerned with integral file, which cannot be used by two
process simultaneously, but I may be mistaken. The same error message arises if I use -np
N*2 switch to mpirun wrapper (where N == number of nodes available).

Additionally, MP2 code doesn't work for me at all. I always get the following message:

Loading P2P interface library... loaded successfully (version 1.8).
Initializing P2P interface... init failed!

All nodes aren't firewalled and have unrestricted rsh access to each other.

The additional trouble may be in 64-bit OS’s installed on our PCs. But most of 32-bit
packages works fine in 32-bit subenvironment, including DFT pcgamess code.

Thanks in advance.

Ivan


Message 589debbbbQn-6587-1290+00.htm, number 220, was posted on Sun Jan 13 '08 at 21:30:27
in reply to c1e90450Lgz-6587-1137-00.htm

Re: Mixed cluster/SMP execution

Pedro Silva
pedros@ufp.pt


Check the FAQ: http://classic.chem.msu.su/gran/gamess/faq.html

(in short: you are out of disk space...)

Pedro S


Message d42eeb47bHW-6587-1327+00.htm, number 221, was posted on Sun Jan 13 '08 at 22:06:54
in reply to c1e90450Lgz-6587-1137-00.htm

Re: Mixed cluster/SMP execution

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. As to updated binaries: please contact me directly to get the latest PC GAMESS builds linked with lam-mpi 7.x.x.

2. As to p2p failure: please attach the output of the "ifconfig -a" to this thread. Please run ifconfig on one of the compute nodes.

3. As to FSF fatal error: please read documentation on the PC GAMESS' command line switches. You can either pass different directories to different PC GAMESS instances manually, or automate the process using -t switch. If this does not help you are likely out of free disk space.

Best regards,
Alex Granovsky



On Sun Jan 13 '08, Ivan Fedyanin wrote
--------------------------------------
>Hi all,
>I'm trying to execute pcgamess on a dual-core AMD64-based linux cluster.

>First of all, I use the following version: [version 7.1 (Tornado), build number 4444].
>Yeah, I do know that this is a beta-release, but it's the one I can find to work with
>lam-mpi 7.x.x.

>DFT-code works pretty well if only one core of the processor is to be used. But as long
>as I put corresponding instruction (cpu=2) in my file, describing the nodes, program
>prints the following error:

>FSF: fatal error no. 0x04000000 in sub PACKREAD on unit 8

>My guess is that this error is concerned with integral file, which cannot be used by two
>process simultaneously, but I may be mistaken. The same error message arises if I use -np
>N*2 switch to mpirun wrapper (where N == number of nodes available).

>Additionally, MP2 code doesn't work for me at all. I always get the following message:

>Loading P2P interface library... loaded successfully (version 1.8).
>Initializing P2P interface... init failed!

>All nodes aren't firewalled and have unrestricted rsh access to each other.

>The additional trouble may be in 64-bit OS’s installed on our PCs. But most of 32-bit
>packages works fine in 32-bit subenvironment, including DFT pcgamess code.

>Thanks in advance.

>Ivan


Message c1e90450Lgz-6587-1338+00.htm, number 222, was posted on Sun Jan 13 '08 at 22:18:08
in reply to 589debbbbQn-6587-1290+00.htm

Re^2: Mixed cluster/SMP execution

Ivan Fedyanin
octy@xrlab.ineos.ac.ru


On Sun Jan 13 '08, Pedro Silva wrote
------------------------------------
>Check the FAQ:
>(in short: you are out of disk space...)
>Pedro S

Well, I think that in my case the problem is different.
In fact, I've got more than 80 GiB of free space available on each node.
My test job is quite small and the maximum size of all files (AOINTS, DICTNRY etc) never
exceeds 500 MiB. Moreover then I run job at one core, I don't get this error message.

After all, I've tried to set FASTF to .f. and got another error:

*ERR* IO-09 system file error - invalid record

but _after_ the program wrote integrals on disk.

Has anyone any success in running pcgamess with mixed SMP/cluster mode?


Message c1e90450Lgz-6587-1359+00.htm, number 223, was posted on Sun Jan 13 '08 at 22:39:53
in reply to d42eeb47bHW-6587-1327+00.htm

Re^2: Mixed cluster/SMP execution

Ivan Fedyanin
octy@xrlab.ineos.ac.ru


Hi,

Ok, thanks, the first problem has been solved. It was also my fault: I used relative but not absolute paths as the parameters for -i and -o switches. It works fine with one cpu but this doesn't work then -t switch is used.

Output of ifconfig -a is attached. This (main) node has two network interfaces, internal is 192.168.1.



This message contains the 2 kb attachment
[ ifconfig.txt ]


Message d42eec98bHW-6588-838+00.htm, number 224, was posted on Mon Jan 14 '08 at 13:57:51
in reply to c1e90450Lgz-6587-1359+00.htm

Re^3: Mixed cluster/SMP execution

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please check if any of these options or their combination help:

 $p2p bind=0 $end
   alternatively
 $p2p bind=1 $end

   and
   
 $p2p incrd(1)=-1 outcrd(1)=-1 $end

Best regards,
Alex Granovsky


On Sun Jan 13 '08, Ivan Fedyanin wrote
--------------------------------------
>Hi,

>Ok, thanks, the first problem has been solved. It was also my fault: I used relative but not absolute paths as the parameters for -i and -o switches. It works fine with one cpu but this doesn't work then -t switch is used.

>Output of ifconfig -a is attached. This (main) node has two network interfaces, internal is 192.168.1.
>
>
>
>


Message c1e90450Lgz-6588-911+00.htm, number 225, was posted on Mon Jan 14 '08 at 15:18:14
in reply to d42eec98bHW-6588-838+00.htm

Re^4: Mixed cluster/SMP execution

Ivan Fedyanin
octy@xrlab.ineos.ac.ru


>Hi,
>please check if any of these options or their combination help:
 $p2p bind=0 $end
   alternatively
 $p2p bind=1 $end

   and
   
 $p2p incrd(1)=-1 outcrd(1)=-1 $end

>Best regards,
>Alex Granovsky

Great.
The combination of bind=1 (or without "bind" at all) and incrd(1)=-1 outcrd(1)=-1 works for me.


Message c3d0f5fe469-6598-1188-00.htm, number 226, was posted on Thu Jan 24 '08 at 19:49:31
How to force PCGAMESS to use other file name then PUNCH?

Igor Shcherbakov
shcherbakov@rsu.ru


Dear PCGAMESS Guru,

Is it possible to force PCGAMESS to use during runtime other filename
for punch file?

Thank you in advance,
Igor Shcherbakov


Message 4df668afbHW-6599-912+00.htm, number 227, was posted on Fri Jan 25 '08 at 15:12:15
in reply to c3d0f5fe469-6598-1188-00.htm

Re: How to force PCGAMESS to use other file name then PUNCH?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. Use -p switch to redirect punch (and other text output files) to the directory where the input file resides

or

2. You can assign names to text output files via environment variables, e.g.:

SET PUNCH=test.pnc

or

export PUNCH=test.pnc

Best regards,
Alex Granovsky




On Thu Jan 24 '08, Igor Shcherbakov wrote
-----------------------------------------
>Dear PCGAMESS Guru,

>Is it possible to force PCGAMESS to use during runtime other filename
>for punch file?

>Thank you in advance,
>Igor Shcherbakov


Message 230aec51jxY-6599-1339-00.htm, number 228, was posted on Fri Jan 25 '08 at 22:20:48
Optimization error regarding the number of translation/rotation modes

Cun Liu
liucun@gmail.com


Dear PC GAMESS users,
I try to optimize an alkali cation benzene complex using HF/3-21G. The optimization failed with following error message. Changing coordinates to z-matrix or rotating molecule could not fix the problems. Can anybody tell me how to get around it? Many thanks!

Best regards,
Cun Liu

*** ERROR! THERE ARE NOT 5 OR 6 TRANS/ROT MODES NUM T/R= 6

THE CAUSE OF THIS IS OFTEN CORRUPTION OF THE HESSIAN
MATRIX AFTER MANY STEPS. THE FIX MAY BE AS SIMPLE AS
RESTARTING THE COMPUTATION, OR MIGHT REQUIRE THAT YOU
CONSIDER CHANGING THE COORDINATES BEING USED.


Message 4df668afbHW-6600-810+00.htm, number 229, was posted on Sat Jan 26 '08 at 13:30:20
in reply to 230aec51jxY-6599-1339-00.htm

Re: Optimization error regarding the number of translation/rotation modes

Alex Granovsky
gran@classic.chem.msu.su


Hi,

try to add: $statpt projct=.f. $end

Best regards,
Alex Granovsky



On Fri Jan 25 '08, Cun Liu wrote
--------------------------------
>Dear PC GAMESS users,
>I try to optimize an alkali cation benzene complex using HF/3-21G. The optimization failed with following error message. Changing coordinates to z-matrix or rotating molecule could not fix the problems. Can anybody tell me how to get around it? Many thanks!

>Best regards,
>Cun Liu

>*** ERROR! THERE ARE NOT 5 OR 6 TRANS/ROT MODES NUM T/R= 6

> THE CAUSE OF THIS IS OFTEN CORRUPTION OF THE HESSIAN
> MATRIX AFTER MANY STEPS. THE FIX MAY BE AS SIMPLE AS
> RESTARTING THE COMPUTATION, OR MIGHT REQUIRE THAT YOU
> CONSIDER CHANGING THE COORDINATES BEING USED.


Message 7c39290d5VG-6604-571-00.htm, number 230, was posted on Wed Jan 30 '08 at 09:31:14
Specific options corresponding to Gaussian03 Opt keyword?

Yangsoo Kim
hit8run@gmail.com


Hellow,

I am trying to calculate the energy and frequency at the B3LYP/6-31G* methods.

Can you recommend $STATPT(OPTTOL, STSTEP, NSTEP, etc...) and $DFT setting corresponding to Opt Freq keywords (default) defined in Gaussian03?

I will appreciate any comments. Thank you in advance...


Yangsoo Kim


Message 80a4ef42ffW-6605-1264-00.htm, number 231, was posted on Thu Jan 31 '08 at 21:04:03
electron density and radial distribution?

esra
esra@gwu.edu


Hello
I am trying to understand the theory behind ESP calculation. I know that to calculate the ESP one needs nuclei effect and electron density effect on the point charge. nuclei effect is easy to calculate. but to calculate the electron density, we need a function to pout on the integral and calculate the area under the curve? what does that function look like? I mean, we need to calculate the effect of point charge on electron density in each volume of the space. but I was wondering what PCGAMESS is using as a electron density function. It should look like a radial distribution function, I guess. but radial distribution changes from s, p,d,f orbitals. Does PCGAMESS taking an average of the radial distribution function and using it? Thank you very much in advance...

Message d4c0f196bHW-6606-967+00.htm, number 232, was posted on Fri Feb 1 '08 at 16:07:38
in reply to 7c39290d5VG-6604-571-00.htm

Re: Specific options corresponding to Gaussian03 Opt keyword?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I would suggest you to leave most of the options at their default values. Changing NSTEP to be e.g., 1000 is a good idea. You can also consider to use DLC coordinates and $statpt method=gdiis geometry optimizer.

Best regards,
Alex Granovsky


On Wed Jan 30 '08, Yangsoo Kim wrote
------------------------------------
>Hellow,

>I am trying to calculate the energy and frequency at the B3LYP/6-31G* methods.

>Can you recommend $STATPT(OPTTOL, STSTEP, NSTEP, etc...) and $DFT setting corresponding to Opt Freq keywords (default) defined in Gaussian03?

>I will appreciate any comments. Thank you in advance...
>
>
>Yangsoo Kim


Message 8072e8f0Uag-6606-1240-00.htm, number 233, was posted on Fri Feb 1 '08 at 20:40:44
generating spin density plot

Mike
rose@chemistry.ucsc.edu


I have difficulty to generate a spin density plot using "cube" command. Is the final spin density mesh in the output, PUNCH, or separate "cube" file? I can't find it to open in gOpenMol or MacMolPlt. My input file starts like this....

$CONTRL SCFTYP=UHF RUNTYP=ENERGY D5=.TRUE. $END
$CONTRL MAXIT=240 ICHARG=-1 MULT=4 DFTTYP=PW91 $END
$SYSTEM TIMLIM=4800 MEMORY=80000000 AOINTS=DIST $END
$BASIS GBASIS=TZV NDFUNC=1 NPFUNC=1 $END
$ELDENS SPIND=.T. $END
$CUBE CUBE=.T. MESH=MEDIUM $END
$P2P P2P=.T. DLB=.T. MXBUF=2048 $END
$GUESS GUESS=MOREAD NORB=575 $END
$DATA

any help is greatly appreciated.


Message 7c39290dcW1-6607-102-00.htm, number 234, was posted on Sat Feb 2 '08 at 01:42:06
Semop lock failed...?

Tae-Yun Park
taeyun0000@gmail.com


Hi,

I'm trying to run a set of 4 jobs on an AMD Phenom 9500 quadcore machine simultaneously under a Linux OS(CentOS5.1). No problem on running a single job, but when 4 jobs are running background on a simultaneous basis, Ive got the following error:

p0_11832: p4_error: OOPS: semop lock failed: -1

More serious problem is, the 4 running jobs do not always produce the corresponding output files. I guess the calculation was done fine, but for some reason, it is not written on the HD. This problem only happens when I tried to run 4 jobs at the same time.

Any suggestion?

Many thanks in advance.

Tae-Yun Park


Message 4df668afbHW-6607-792+00.htm, number 235, was posted on Sat Feb 2 '08 at 13:11:58
in reply to 7c39290dcW1-6607-102-00.htm

Re: Semop lock failed...?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you simply need to use cleanipcs or ipcrm to remove old dead sysv semaphores from your system. Note, if you are running PC GAMESS not in parallel, you can reduce the number of required semaphores by using -nompi command line switch.

Best regards,
Alex Granovsky


On Sat Feb 2 '08, Tae-Yun Park wrote
------------------------------------
>Hi,

>I'm trying to run a set of 4 jobs on an AMD Phenom 9500 quadcore machine simultaneously under a Linux OS(CentOS5.1). No problem on running a single job, but when 4 jobs are running background on a simultaneous basis, Ive got the following error:

>p0_11832: p4_error: OOPS: semop lock failed: -1

>More serious problem is, the 4 running jobs do not always produce the corresponding output files. I guess the calculation was done fine, but for some reason, it is not written on the HD. This problem only happens when I tried to run 4 jobs at the same time.

>Any suggestion?

>Many thanks in advance.

>Tae-Yun Park
>


Message 4df668afbHW-6607-797+00.htm, number 236, was posted on Sat Feb 2 '08 at 13:17:31
in reply to 80a4ef42ffW-6605-1264-00.htm

Re: electron density and radial distribution?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

having the first-order density matrix D, the electron density is defined as pho(r)=sum(p,q)Dpq*p(r)*q(r)where Dpq is the matrix element of density matrix, and p(r) and q(r) are basis functions. Then the problem reduces to the calculation of ESP of known electronic density.

Best regards,
Alex Granovsky


On Thu Jan 31 '08, esra wrote
-----------------------------
>Hello
>I am trying to understand the theory behind ESP calculation. I know that to calculate the ESP one needs nuclei effect and electron density effect on the point charge. nuclei effect is easy to calculate. but to calculate the electron density, we need a function to pout on the integral and calculate the area under the curve? what does that function look like? I mean, we need to calculate the effect of point charge on electron density in each volume of the space. but I was wondering what PCGAMESS is using as a electron density function. It should look like a radial distribution function, I guess. but radial distribution changes from s, p,d,f orbitals. Does PCGAMESS taking an average of the radial distribution function and using it? Thank you very much in advance...


Message 4df668afbHW-6607-801+00.htm, number 237, was posted on Sat Feb 2 '08 at 13:20:49
in reply to 8072e8f0Uag-6606-1240-00.htm

Re: generating spin density plot

Alex Granovsky
gran@classic.chem.msu.su


Hi,

try to search for $cube group in the punch file. If you are using gOpenMol, use gamess2plt utility to convert text $cube groups of PUNCH to binary gOpenMol format.

Best regards,
Alex Granovsky


On Fri Feb 1 '08, Mike wrote
----------------------------
>I have difficulty to generate a spin density plot using "cube" command. Is the final spin density mesh in the output, PUNCH, or separate "cube" file? I can't find it to open in gOpenMol or MacMolPlt. My input file starts like this....

> $CONTRL SCFTYP=UHF RUNTYP=ENERGY D5=.TRUE. $END
> $CONTRL MAXIT=240 ICHARG=-1 MULT=4 DFTTYP=PW91 $END
> $SYSTEM TIMLIM=4800 MEMORY=80000000 AOINTS=DIST $END
> $BASIS GBASIS=TZV NDFUNC=1 NPFUNC=1 $END
> $ELDENS SPIND=.T. $END
> $CUBE CUBE=.T. MESH=MEDIUM $END
> $P2P P2P=.T. DLB=.T. MXBUF=2048 $END
> $GUESS GUESS=MOREAD NORB=575 $END
> $DATA

>any help is greatly appreciated.
>


Message 7c39290dcW1-6610-656-00.htm, number 238, was posted on Tue Feb 5 '08 at 10:56:56
input & mpirun command of benchmark results ?

Tae-Yun Park
taeyun0000@gmail.com


Hi,

I'm enjoying the performance of PCGAMESS, and Im highly appreciate the effort to develop this excellent QM package.

I wonder if I could get the input file and mpirun commands that produced the benchmark results Test1 shown in http://classic.chem.msu.su/gran/gamess/barcelona.html. We are currently trying to test a similar machine, but only one CPU core is running at the moment. I guess theres something missing in my input & mpirun command settings.

If there is any additional setting related ti mpirun or PCGAMESS input that should be applied to get the similar results of 'Test 1', I would appreciate to be informed on those, too.

I guess the benchmark results in Test 1 are obtained from a single machine with two quad core CPUs (8 CPU cores), not from 8 machines with 16 quad core CPUs (64 CPU cores). Please comment if this is not the case.

Many thanks in advance.

Tae-Yun Park


Message 4df668afbHW-6610-751+00.htm, number 239, was posted on Tue Feb 5 '08 at 12:30:58
in reply to 7c39290dcW1-6610-656-00.htm

Re: input & mpirun command of benchmark results ?

Alex Granovsky
gran@classic.chem.msu.su


Hi,


>I wonder if I could get the input file and mpirun commands that produced the benchmark results Test1 shown in http://classic.chem.msu.su/gran/gamess/barcelona.htmlhttp://classic.chem.msu.su/gran/gamess/barcelona.html. We are currently trying to test a similar machine, but only one CPU core is running at the moment. I guess theres something missing in my input & mpirun command settings.

>If there is any additional setting related ti mpirun or PCGAMESS input that should be applied to get the similar results of 'Test 1', I would appreciate to be informed on those, too.


I can provide you by the input file and instructions on how to run it in parallel. However, the exact answer depends on the OS and MPI version you are using. Please provide more information on these.


>I guess the benchmark results in Test 1 are obtained from a single machine with two quad core CPUs (8 CPU cores),

Yes they are.

Best regards,
Alex Granovsky


Message 7c39290dcW1-6610-807+00.htm, number 240, was posted on Tue Feb 5 '08 at 13:27:22
in reply to 4df668afbHW-6610-751+00.htm

Re^2: input & mpirun command of benchmark results ?

Tae-Yun Park
taeyun0000@gmail.com


Dear Dr. Granovsky:

Many thanks for your prompt reply and kind help.

Here are the information on the OS & the version of MPICH:

Machine 1
OS: AMD Phenom9500 under CentOS 5.1 (kernel2.6.24.9)

Machine 2
OS: Core2Qquad Q6600 under CentOS 5.0 (kernel2.6.22.9)

MPICH version for both machines: 1.2.7p1

>I can provide you by the input file and instructions on how to run it in parallel. However, the exact answer depends on the OS and MPI version you are using. Please provide more information on these.
>

Of course, I can always change & reinstall the OS and MPICH as recommended for better performance of PCGAMESS mpiruns.

Thanks again,

Sincerely,

Tae-Yun Park


Message 4df668afbHW-6610-901+00.htm, number 241, was posted on Tue Feb 5 '08 at 15:01:58
in reply to 7c39290dcW1-6610-807+00.htm

Re^3: input & mpirun command of benchmark results ?

Alex Granovsky
gran@classic.chem.msu.su


Dear Dr. Tae-Yun Park,

attached to my post please find the compressed input file.

You should run it, e.g.:

mpirun -np 4 /home/gran/pcgamess -r -f -p -i /home/gran/test1.inp -o /home/gran/test1.out -ex /home/gran -t /tmp/pcg

Here -t /tmp/pcg points to the scratch directory, while -ex /home/gran points to the place where PC GAMESS runtime extensions (fastdiag.ex and pcgp2p.ex) reside. For more information, please consult with the description of the PC GAMESS command line options.

In the case mpich is not properly configured on your computer, you can always run parallel job manually. For example,

/home/gran/pcgamess -r -f -p -i /home/gran/test1.inp -o /home/gran/test1.out -ex /home/gran -t /tmp/pcg -p4pg /home/gran/pg4

will start four processes in parallel provided that the procgrp file "home/gran/pg4" contains the following line:

local 3


Best regards,
Alex Granovsky









On Tue Feb 5 '08, Tae-Yun Park wrote
------------------------------------
>Dear Dr. Granovsky:

>Many thanks for your prompt reply and kind help.

>Here are the information on the OS & the version of MPICH:

>Machine 1
>OS: AMD Phenom9500 under CentOS 5.1 (kernel2.6.24.9)

>Machine 2
>OS: Core2Qquad Q6600 under CentOS 5.0 (kernel2.6.22.9)

>MPICH version for both machines: 1.2.7p1

>>I can provide you by the input file and instructions on how to run it in parallel. However, the exact answer depends on the OS and MPI version you are using. Please provide more information on these.
>>

>Of course, I can always change & reinstall the OS and MPICH as recommended for better performance of PCGAMESS mpiruns.

>Thanks again,

>Sincerely,

>Tae-Yun Park

This message contains the 2 kb attachment
[ test1.zip ] Benchmark 1 input file.


Message 5c714642PMq-6610-1105-00.htm, number 242, was posted on Tue Feb 5 '08 at 18:26:15
PC Gamess with SLURM

Vitaliy Gorbenko
vgorbenko@ukr.net


Hi,
I'm runnig the single PC GAMESS job on cluster with SLURM system without any problems. Can I run the parallel jobs with SLURM?

Many thanks in advance.

Vitaliy Gorbenko


Message 4df668afbHW-6610-1164+00.htm, number 243, was posted on Tue Feb 5 '08 at 19:24:25
in reply to 5c714642PMq-6610-1105-00.htm

Re: PC Gamess with SLURM

Alex Granovsky
gran@classic.chem.msu.su


Hi,

we never did that but I believe it should be possible to do this exactly as with any other MPICH program.

Best regards,
Alex Granovsky


On Tue Feb 5 '08, Vitaliy Gorbenko wrote
----------------------------------------
>Hi,
>I'm runnig the single PC GAMESS job on cluster with SLURM system without any problems. Can I run the parallel jobs with SLURM?

>Many thanks in advance.

>Vitaliy Gorbenko


Message 935b012dcvZ-6610-1220-00.htm, number 244, was posted on Tue Feb 5 '08 at 20:20:33
NBO 5.0

Zoran Matovic
zmatovic@kg.ac.yu


Hi PCGAMESS users,

Alex said:
"finally we have the beta version of PC GAMESS to test with NBO 5.0 incorporated into it"

Can we try beta version?

best wishes for all of you
zoran


Message 3a48de3acW1-6610-1316+00.htm, number 245, was posted on Tue Feb 5 '08 at 21:56:15
in reply to 4df668afbHW-6610-901+00.htm

Re^4: Many thanks

Tae-Yun Park
taeyun0000@gmail.com


Dear Dr. Granovsky,

Many thanks for your reply. Your answer was of great help, and our maches worked just fine!

Sincerely,

Tae-Yun Park


Message 505171cb7WL-6611-746-00.htm, number 246, was posted on Wed Feb 6 '08 at 12:25:46
Uncompressing PCGAMESS on Ubuntu (and probably other distros too)

Fernando Muniz
fernando.inbox@gmail.com


I have unzipped the pcgamess_linux_mpich_p4.zip file in my home directory using the first password, getting a single windows style executable file. I tried to launch it and it didn't give any output. I asked Mr. Granovsky to give me a hunch and so I found out that this .exe file was a self extracting executable. I tried Ark, Archive Manager, orange, and unzip on it and it didn't recognize the file. In this discussion I discovered that the .exe file was in rar format.

Then I tried then following and it worked (asking for the second password):
"unrar -e pcg710.exe"


Message 4df668afbHW-6611-972+00.htm, number 247, was posted on Wed Feb 6 '08 at 16:12:18
in reply to 935b012dcvZ-6610-1220-00.htm

Re: NBO 5.0

Alex Granovsky
gran@classic.chem.msu.su


Hi,

if you already have valid PC GAMESS' NBO license, please contact me directly and specify for which OS/CPU/MPI combinations you need the updated binaries.

Best regards,
Alex Granovsky


On Tue Feb 5 '08, Zoran Matovic wrote
-------------------------------------
>Hi PCGAMESS users,

>Alex said:
>"finally we have the beta version of PC GAMESS to test with NBO 5.0 incorporated into it"

>Can we try beta version?

>best wishes for all of you
>zoran


Message 4df668afbHW-6611-978+00.htm, number 248, was posted on Wed Feb 6 '08 at 16:18:18
in reply to 505171cb7WL-6611-746-00.htm

Re: Uncompressing PCGAMESS on Ubuntu (and probably other distros too)

Alex Granovsky
gran@classic.chem.msu.su


Hi,

note that the pcg710.exe is the valid Linux ELF executable file (not the Windows one despite its .exe suffix!), namely it's a rar self-extracting archive. Unfortunately due to unknown reasons the selfextractor we used does not work under some of the modern Linux distributions, e.g., Ubuntu and SuSE.

Best regards,
Alex

On Wed Feb 6 '08, Fernando Muniz wrote
--------------------------------------
>I have unzipped the pcgamess_linux_mpich_p4.zip file in my home directory using the first password, getting a single windows style executable file. I tried to launch it and it didn't give any output. I asked Mr. Granovsky to give me a hunch and so I found out that this .exe file was a self extracting executable. I tried Ark, Archive Manager, orange, and unzip on it and it didn't recognize the file. In this discussion I discovered that the .exe file was in rar format.

>Then I tried then following and it worked (asking for the second password):
>"unrar -e pcg710.exe"


Message 93868d6dSol-6611-1087-00.htm, number 249, was posted on Wed Feb 6 '08 at 18:07:20
spin components of the MP2 correlation energy

Jozsef Csontos
jcsontos.lists@gmail.com


Hi,

First of all, I'm a very novice user of PC-GAMESS; I just started using a few days ago.
This was just for the record if my questions are going to be too evident. (More precisely the answers for those.)

Yesterday, I was browsing the archive and I found out that if you want to get the singlet and triplet spin contributions to the MP2 correlation energy, you need to do an MP3 calculation. This was in a message from the end of 2006. Is this still a valid statement or this was changed in the code since then?

It seems to me that the analytical MP2 frequency calculation hasn't been implemented yet, has it?

My third and definitely lame question is: the web page posting is the only way to send messages to this discussion forum?

Thank you,

Jozsef


Message 4df668afbHW-6612-828+00.htm, number 250, was posted on Thu Feb 7 '08 at 13:48:38
in reply to 93868d6dSol-6611-1087-00.htm

Re: spin components of the MP2 correlation energy

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>First of all, I'm a very novice user of PC-GAMESS; I just started using a few days ago.

welcome to the club.


>Yesterday, I was browsing the archive and I found out that if you want to get the singlet and triplet spin contributions to the MP2 correlation energy, you need to do an MP3 calculation. This was in a message from the end of 2006. Is this still a valid statement or this was changed in the code since then?

The situation was not changed since then, however it seems
I overlooked another, more cheaper way how to get them.
More precisely, you can run UHF (note not RHF!) MP2 calculations
for singlet state with $mp2 method=3 $end. Then you can easily
extract singlet and triplet pair contributions to E(2) by
multiplying printed value of

  alpha alpha alpha alpha

contribution to E(2) by the factor of three.

Method=3 is not very appropriate for large scale jobs, however,
if there is a considerable interest to get the separate contributions
of singlet and triplet pairs to E(2), we can consider
implementation of this feature in the standard closed-shell
MP2 code.

The reason why this is not implemented is that these contributions
are seldom used separately, unless you are going to scale them by
the different factors.


>It seems to me that the analytical MP2 frequency calculation hasn't been implemented yet, has it?

Yes they have not yet implemented.

>My third and definitely lame question is: the web page posting is the only way to send messages to this discussion forum?

Yes this is the only way.


Best regards,
Alex Granovsky


Message c6b48310SZP-6613-2-00.htm, number 251, was posted on Fri Feb 8 '08 at 00:02:43
HESSIAN restart

Slawomir Janicki
slawomir.janicki@comcast.net


Hi,

I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
**** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
GOT JVIB,JATOM,JCOORD= 0 0 0

Do you know how to fix that?

SJ


Message d4c0f196bHW-6613-1003+00.htm, number 252, was posted on Fri Feb 8 '08 at 16:43:02
in reply to c6b48310SZP-6613-2-00.htm

Re: HESSIAN restart

Alexei Popov
alexei.a.popov@gmail.com


Hi,

maybe you copied it twice from ircdata by mistake?

Regards,
Alexei

On Fri Feb 8 '08, Slawomir Janicki wrote
----------------------------------------
>Hi,

>I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

> IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
> IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
> **** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
> GOT JVIB,JATOM,JCOORD= 0 0 0

>Do you know how to fix that?

>SJ


Message c6b48310SZP-6613-1017+00.htm, number 253, was posted on Fri Feb 8 '08 at 16:57:30
in reply to d4c0f196bHW-6613-1003+00.htm

Re^2: HESSIAN restart

Slawomir Janicki
slawomir.janicki@comcast.net


Nope, I checked again and there is only one copy. PC GAMESS clearly wants an energy that has not been calculated.

SJ

On Fri Feb 8 '08, Alexei Popov wrote
------------------------------------
>Hi,

>maybe you copied it twice from ircdata by mistake?

>Regards,
>Alexei

>On Fri Feb 8 '08, Slawomir Janicki wrote
>----------------------------------------
>>Hi,

>>I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

>> IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
>> IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
>> **** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
>> GOT JVIB,JATOM,JCOORD= 0 0 0

>>Do you know how to fix that?

>>SJ


Message d4c0f196bHW-6613-1026+00.htm, number 254, was posted on Fri Feb 8 '08 at 17:06:39
in reply to c6b48310SZP-6613-1017+00.htm

Re^3: HESSIAN restart

Alex Granovsky
gran@classic.chem.msu.su


Hi,

could you please send me your input file?

Cheers,
Alex


On Fri Feb 8 '08, Slawomir Janicki wrote
----------------------------------------
>Nope, I checked again and there is only one copy. PC GAMESS clearly wants an energy that has not been calculated.

>SJ

>On Fri Feb 8 '08, Alexei Popov wrote
>------------------------------------
>>Hi,

>>maybe you copied it twice from ircdata by mistake?

>>Regards,
>>Alexei

>>On Fri Feb 8 '08, Slawomir Janicki wrote
>>----------------------------------------
>>>Hi,

>>>I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

>>> IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
>>> IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
>>> **** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
>>> GOT JVIB,JATOM,JCOORD= 0 0 0

>>>Do you know how to fix that?

>>>SJ


Message 4df668afbHW-6613-1220+00.htm, number 255, was posted on Fri Feb 8 '08 at 20:20:08
in reply to c6b48310SZP-6613-1017+00.htm

Re^3: HESSIAN restart

Alexei Popov
alexei.a.popov@gmail.com


Hi,

check if you entered

 $end

not

$end

-AP


On Fri Feb 8 '08, Slawomir Janicki wrote
----------------------------------------
>Nope, I checked again and there is only one copy. PC GAMESS clearly wants an energy that has not been calculated.

>SJ

>On Fri Feb 8 '08, Alexei Popov wrote
>------------------------------------
>>Hi,

>>maybe you copied it twice from ircdata by mistake?

>>Regards,
>>Alexei

>>On Fri Feb 8 '08, Slawomir Janicki wrote
>>----------------------------------------
>>>Hi,

>>>I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

>>> IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
>>> IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
>>> **** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
>>> GOT JVIB,JATOM,JCOORD= 0 0 0

>>>Do you know how to fix that?

>>>SJ


Message c6b48310SZP-6614-15+00.htm, number 256, was posted on Sat Feb 9 '08 at 00:14:43
in reply to 4df668afbHW-6613-1220+00.htm

Re^4: HESSIAN restart

Slawomir Janicki
slawomir.janicki@comcast.net


Hi Alexei,

That was it, thanks for spotting the problem. And I used to like Fortran....

SJ

On Fri Feb 8 '08, Alexei Popov wrote
------------------------------------
>Hi,

>check if you entered

>

 $end

>not

>

$end

>-AP
>
>
>On Fri Feb 8 '08, Slawomir Janicki wrote
>----------------------------------------
>>Nope, I checked again and there is only one copy. PC GAMESS clearly wants an energy that has not been calculated.

>>SJ

>>On Fri Feb 8 '08, Alexei Popov wrote
>>------------------------------------
>>>Hi,

>>>maybe you copied it twice from ircdata by mistake?

>>>Regards,
>>>Alexei

>>>On Fri Feb 8 '08, Slawomir Janicki wrote
>>>----------------------------------------
>>>>Hi,

>>>>I am trying to restart a numerical HESSIAN job. According to the manual I should include the $VIB group from the incomplete IRCDATA file and add $END at the end. When I do that I get the following error:

>>>> IVIB= 2 IATOM= 2 ICOORD= 3 E= -330.8873973598
>>>> IVIB= 2 IATOM= 3 ICOORD= 1 E= -330.8873976826
>>>> **** ERROR, EXPECTING IVIB,IATOM,ICOORD= 2 3 2
>>>> GOT JVIB,JATOM,JCOORD= 0 0 0

>>>>Do you know how to fix that?

>>>>SJ


Message 5c715b0aPMq-6615-1199-00.htm, number 257, was posted on Sun Feb 10 '08 at 19:58:39
How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


Hi,
I'm using SSH on the cluster nodes. SSH has been configured for passwordless using. For parallel runing of pcgamess I used in comandline
-p4pg procpg
with next lines in file
n01 3
n02 3
The massege "Permission denied, please try again" has been obtained in the answer. Is rsh using in PCGAMESS as default? How to change it to SSH?

Thank you,
Vitaliy Gorbenko


Message 4df668afbHW-6615-1209+00.htm, number 258, was posted on Sun Feb 10 '08 at 20:08:59
in reply to 5c715b0aPMq-6615-1199-00.htm

Re: How to change rsh to ssh?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

export P4_RSHCOMMAND=ssh

Regards,
Alex Granovsky



On Sun Feb 10 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>Hi,
>I'm using SSH on the cluster nodes. SSH has been configured for passwordless using. For parallel runing of pcgamess I used in comandline
> -p4pg procpg
>with next lines in file
> n01 3
> n02 3
>The massege "Permission denied, please try again" has been obtained in the answer. Is rsh using in PCGAMESS as default? How to change it to SSH?

>Thank you,
>Vitaliy Gorbenko
>


Message 5c715b0aPMq-6615-1307+00.htm, number 259, was posted on Sun Feb 10 '08 at 21:47:05
in reply to 4df668afbHW-6615-1209+00.htm

Re^2: How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


On Sun Feb 10 '08, Alex Granovsky wrote
---------------------------------------
>Hi,

>export P4_RSHCOMMAND=ssh

>Regards,
>Alex Granovsky
>
>
>

Now I've tried this but result is remain negative. I look the environment and found P4_RSHCOMMAND=ssh.
Can you suggest anything me?

>>Thank you,
>>Vitaliy Gorbenko
>>


Message 5c714a21PMq-6616-686+00.htm, number 260, was posted on Mon Feb 11 '08 at 11:26:19
in reply to 5c715b0aPMq-6615-1307+00.htm

Re^3: How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


Hi!

What is my mistake if I receiving the message

p0_10425: p4_error: create_procgroup: getpwuid failed: 0

Help me pls.

Vitaliy Gorbenko


Message 4df668afbHW-6616-772+00.htm, number 261, was posted on Mon Feb 11 '08 at 12:52:39
in reply to 5c714a21PMq-6616-686+00.htm

Re^4: How to change rsh to ssh?

Alexei Popov
alexei.a.popov@gmail.com


Hi,

http://classic.chem.msu.su/gran/gamess/introlinux.html

Regards,
Alexei

On Mon Feb 11 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>Hi!

>What is my mistake if I receiving the message

>p0_10425: p4_error: create_procgroup: getpwuid failed: 0

>Help me pls.

>Vitaliy Gorbenko


Message 8afd7dbbaKC-6617-118-00.htm, number 262, was posted on Tue Feb 12 '08 at 01:58:02
relativity ,

Richard
bonarlaw@liv.ac.uk


Are there any plans to introduce the $RELWFN group? Gamess-US seems slow and prone to DAF write errors with the basis sets used for these calculations.

Richard


Message d4c0f196bHW-6617-911+00.htm, number 263, was posted on Tue Feb 12 '08 at 15:10:59
in reply to 8afd7dbbaKC-6617-118-00.htm

Re: relativity ,

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there are no such plans at moment.
However, we are now considering the possibility of our own implementation of Spin-Orbit-DFT similar to that implemented in the NWCHEM.

Best regards,
Alex Granovsky


On Tue Feb 12 '08, Richard wrote
--------------------------------
>Are there any plans to introduce the $RELWFN group? Gamess-US seems slow and prone to DAF write errors with the basis sets used for these calculations.

>Richard


Message 7c39290dcW1-6621-732-00.htm, number 264, was posted on Sat Feb 16 '08 at 12:12:24
Automizing OPT & FREQ followed by Raman calculation?

Tae-Yun Park
taeyun0000@gmail.com


Dear PCGAMESS users,

I wonder if anyone of you performs Raman calculations with PCGAMESS. In this case, to my best knowledge at this time, I need to perform two steps separately, i.e. OPT & FREQ as a first step and then RAMAN as a second step.

Especially, I need to manually copy the energy & Hessian information obtained from the punch file of the first step and then paste into the input file of the second step to do the Raman calculation.

I wonder if there is any way to automize the first step and the second step, which can avoid the manual copy and paste process.

Any comment would be greatly appreciated.

Many thanks in advance.

Tae-Yun Park


Message 7c39290d5VG-6621-795-00.htm, number 265, was posted on Sat Feb 16 '08 at 13:15:50
How to Calculate the frequency-dependent polarizability & optical rotation?

Yangsoo Kim
hit8run@gmail.com


Hi,

I want to calculate the frequency-dependent polarizabilities/hyperpolarizabilities(Gaussian Keywords: Polar CPHF=RdFreq) and Optical rotations(Gaussian Keyworkds: Polar=OptRot CPHF=RdFreq) at the specified frequency.
Is it possible in PC-Gamess? If yes, give me the specific keywords for these properties, please.

Many thanks...

Yangsoo Kim


Message 4df668afbHW-6621-890+00.htm, number 266, was posted on Sat Feb 16 '08 at 14:50:07
in reply to 7c39290d5VG-6621-795-00.htm

Re: How to Calculate the frequency-dependent polarizability & optical rotation?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

for RHF frequency-dependent polarizabilities, you should use runtyp=tdhf. Please read the corresponding sections of the PC GAMESS manual. Starting from the PC GAMESS v. 7.0, TDHF and DFT code were extended to allow calculation of static and dynamic (hyper)polarizabilities using runtyp=tdhf as well. For DFT, alpha values are exact, while beta and gamma are only approximate at present, as second-order (and higher) exchange-correlation kernels are not properly taken into account.

It is not possible to calculate optical rotations with PC GAMESS at present.

Best regards,
Alex Granovsky

On Sat Feb 16 '08, Yangsoo Kim wrote
------------------------------------
>Hi,

>I want to calculate the frequency-dependent polarizabilities/hyperpolarizabilities(Gaussian Keywords: Polar CPHF=RdFreq) and Optical rotations(Gaussian Keyworkds: Polar=OptRot CPHF=RdFreq) at the specified frequency.
>Is it possible in PC-Gamess? If yes, give me the specific keywords for these properties, please.

>Many thanks...

>Yangsoo Kim

>


Message 4df668afbHW-6621-892+00.htm, number 267, was posted on Sat Feb 16 '08 at 14:52:03
in reply to 7c39290dcW1-6621-732-00.htm

Re: Automizing OPT & FREQ followed by Raman calculation?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

try to use JOB macrolanguage (available in the "Downloads" section) to automate the process.

Best regards,
Alex Granovsky


On Sat Feb 16 '08, Tae-Yun Park wrote
-------------------------------------
>Dear PCGAMESS users,

>I wonder if anyone of you performs Raman calculations with PCGAMESS. In this case, to my best knowledge at this time, I need to perform two steps separately, i.e. OPT & FREQ as a first step and then RAMAN as a second step.

>Especially, I need to manually copy the energy & Hessian information obtained from the punch file of the first step and then paste into the input file of the second step to do the Raman calculation.

>I wonder if there is any way to automize the first step and the second step, which can avoid the manual copy and paste process.

>Any comment would be greatly appreciated.

>Many thanks in advance.

>Tae-Yun Park
>


Message 5c715226PMq-6622-33+00.htm, number 268, was posted on Sun Feb 17 '08 at 00:32:53
in reply to 4df668afbHW-6616-772+00.htm

Re^5: How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


On Mon Feb 11 '08, Alexei Popov wrote
-------------------------------------
>Hi,

>http://classic.chem.msu.su/gran/gamess/introlinux.htmlhttp://classic.chem.msu.su/gran/gamess/introlinux.html

>Regards,
>Alexei

I did try the next commandline
ssh -t -t $HOSTNAME /home/gorbenko/pcgamess715/pcgamess "$@" ....

but the message has been recieved

tcgetattr: Invalid argument
n02: Connection refused

Next I ran PCGAMESS and wrote the list of processes. Among them I found such line

5411 pts/5 SN+ 0:00 /usr/bin/rsh n02 -l gorbenko -n /home/gorbenko/pcgamess715/pcgamess n01 56969 \-p4amslave \-p4yourname n02 \-p4rmrank 1

May be "Connection refused" is caused the using rsh-protocol?
What can I do?
Help me.






Message 4df668afbHW-6622-1124+00.htm, number 269, was posted on Sun Feb 17 '08 at 18:44:16
in reply to 5c715226PMq-6622-33+00.htm

Re^6: How to change rsh to ssh?

Alexei Popov
alexei.a.popov@gmail.com


Hi,

first, you are using wrong format for procgrp file.
The correct format to start, e.g., PC GAMESS on 16 dual-CPU (or just dual core) boxes is:

local 1
node1 2 /home/avb/715/pcgamess
node2 2 /home/avb/715/pcgamess
node3 2 /home/avb/715/pcgamess
node4 2 /home/avb/715/pcgamess
node5 2 /home/avb/715/pcgamess
node6 2 /home/avb/715/pcgamess
node7 2 /home/avb/715/pcgamess
node8 2 /home/avb/715/pcgamess
node9 2 /home/avb/715/pcgamess
node10 2 /home/avb/715/pcgamess
node11 2 /home/avb/715/pcgamess
node12 2 /home/avb/715/pcgamess
node13 2 /home/avb/715/pcgamess
node14 2 /home/avb/715/pcgamess
node15 2 /home/avb/715/pcgamess

You need to provide the path to executable in the procgrp file!

Second, if statically linked binaries do not work for you, try dynamically-linked ones.

Finally, you should export your P4_RSHCOMMAND environment variable!

Regards,
Alexei




On Sun Feb 17 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>On Mon Feb 11 '08, Alexei Popov wrote
>-------------------------------------
>>Hi,

>>http://classic.chem.msu.su/gran/gamess/introlinux.htmlhttp://classic.chem.msu.su/gran/gamess/introlinux.htmlhttp://classic.chem.msu.su/gran/gamess/introlinux.htmlhttp://classic.chem.msu.su/gran/gamess/introlinux.html

>>Regards,
>>Alexei

>I did try the next commandline
>ssh -t -t $HOSTNAME /home/gorbenko/pcgamess715/pcgamess "$@" ....

>but the message has been recieved

>tcgetattr: Invalid argument
>n02: Connection refused

>Next I ran PCGAMESS and wrote the list of processes. Among them I found such line

>5411 pts/5 SN+ 0:00 /usr/bin/rsh n02 -l gorbenko -n /home/gorbenko/pcgamess715/pcgamess n01 56969 \-p4amslave \-p4yourname n02 \-p4rmrank 1

>May be "Connection refused" is caused the using rsh-protocol?
>What can I do?
>Help me.
>
>
>
>
>
>


Message 5c714e7ePMq-6622-1391+00.htm, number 270, was posted on Sun Feb 17 '08 at 23:12:35
in reply to 4df668afbHW-6622-1124+00.htm

Re^7: How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


On Sun Feb 17 '08, Alexei Popov wrote
-------------------------------------
>Hi,

>first, you are using wrong format for procgrp file.
>The correct format to start, e.g., PC GAMESS on 16 dual-CPU (or just dual core) boxes is:

>local 1
>node1 2 /home/avb/715/pcgamess
>node2 2 /home/avb/715/pcgamess
>node3 2 /home/avb/715/pcgamess
>node4 2 /home/avb/715/pcgamess
>node5 2 /home/avb/715/pcgamess
>node6 2 /home/avb/715/pcgamess
>node7 2 /home/avb/715/pcgamess
>node8 2 /home/avb/715/pcgamess
>node9 2 /home/avb/715/pcgamess
>node10 2 /home/avb/715/pcgamess
>node11 2 /home/avb/715/pcgamess
>node12 2 /home/avb/715/pcgamess
>node13 2 /home/avb/715/pcgamess
>node14 2 /home/avb/715/pcgamess
>node15 2 /home/avb/715/pcgamess

>You need to provide the path to executable in the procgrp file!

>Second, if statically linked binaries do not work for you, try dynamically-linked ones.

>Finally, you should export your P4_RSHCOMMAND environment variable!

>Regards,
>Alexei
>
Hi,

first of all I've ran PCGAMESS on the every nodes in sequentially.

Next, I've ran PCGAMESS in SMP parallel mode as written in

http://classic.chem.msu.su/gran/gamess/comm_line.html
and
http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-6610-901+00.htm
without any problems. :)
Some jobs has been solved successful.

Next, I've included in .bash_profile the line

export P4_RSHCOMMAND=ssh

and reloging in the system.

As recomended in

http://classic.chem.msu.su/gran/gamess/linux_mpich.html

in procgrp-file I wrote lines

local 0
n02 2 /home/gorbenko/pcgamess715/pcgamess

and have tried to run PCGAMESS in parallel mode.

In answer I recived message

p0_10425: p4_error: create_procgroup: getpwuid failed: 0

You gave me recomandation to read

http://classic.chem.msu.su/gran/gamess/introlinux.html

I read and made

http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C35c715226PMq-6622-33+00.htm

I'm using procgrp-file like to your one (see above). At running PCGAMESS I recived in answer

n02: Connection refused

with line in proccesses list

5411 pts/5 SN+ 0:00 /usr/bin/rsh n02 -l gorbenko -n /home/gorbenko/pcgamess715/pcgamess n01 56969 \-p4amslave \-p4yourname n02 \-p4rmrank 1

What is wrong?

Regards,

Vitaliy Gorbenko








Message 4df668afbHW-6623-854+00.htm, number 271, was posted on Mon Feb 18 '08 at 14:14:10
in reply to 5c714e7ePMq-6622-1391+00.htm

Re^8: How to change rsh to ssh?

Alexei Popov
alexei.a.popov@gmail.com


Hi,

are you trying to run PC GAMESS in parallel manually or using some queue/batch system? In the last case, you should modify .bashrc, not .bash_profile

Regards,
Alexei


Message 7c39290dcW1-6623-920-00.htm, number 272, was posted on Mon Feb 18 '08 at 15:20:27
PCGAMESS default OPT options compared to G03?

Tae-Yun Park
taeyun0000@gmail.com


Dear PCGAMESS users,

While Ive been enjoying the performance of PCGAMESS recently, I found another factor that the performance of PCGAMESS could possibly be compared unfairly if one just uses default settings for geometry optimization. The default settings in PCGAMESS are,
OPTTOL = 0.00010
RMS Force = OPTTOL/3 = 0.00003333
RMAX = 0.1
RMIN = 0.0015

In the case of G03, the default settings are,
Max. Force = 0.00045
RMS Force = 0.00030
Max. Displacement = 0.00180
RMS Displacement = 0.00120

If OPTTOL and Max. Force have got the same meaning, it looks to me that PCGAMESS settings are more severe, which may result in more iterations and calculation time but more accurate geometry. Same for the case of RMS Force setting. I wonder if this is a correct view.

If indeed the default settings are not really equivalent, I wonder if the PCGAMESS OPTTOL can be set to 0.00045 to make the similar conditions compared to the case of G03. This makes PCGAMESS RMS Force = OPTTOL/3 = 0.00015, which is still smaller than G03. If it is so, what about the RMAX and RMIN? What values could to be used to have the similar conditions to G03 default settings?

Finally, if indeed the PCGAMESS default settings are more sever compared to those of G03, is there any specific reason for that?

Any comments would be highly appreciated.

Many thanks,

Tae-Yun


Message 4df668afbHW-6625-1021+00.htm, number 273, was posted on Wed Feb 20 '08 at 17:01:14
in reply to 7c39290dcW1-6623-920-00.htm

Re: PCGAMESS default OPT options compared to G03?

Alex Granovsky
gran@classic.chem.msu.su


Dear Tae-Yun,

actually, PC GAMESS uses only two criteria during geometry optimization - namely, maximum gradient and RMS gradient. The reason of this is that maximum component of displacement and RMS displacement are primary the property of geometry optimization routine used, not the PES itself (while gradients are the property of PES only).

It is always hard to compare two different programs, especially PC GAMESS with Gaussian as we are banned by Gaussian and do not have access to it. However, according to multiple reports of our users, PC GAMESS by default indeed has more strict criteria for geometry optimization. Basically, this is because in most cases the allowed RMS force is the strongest component of two criteria used by PC GAMESS, and they differ considerably between PC GAMESS and Gaussian.

The reason why PC GAMESS uses stronger criteria than Gaussian is that in the case of flat PES, you may find significant difference between geometries optimized at OPTTOL=1d-4 and, say, OPTTOL=5d-4. In general, the default set of parameters used by PC GAMESS seems to be reasonable for almost all cases. However, for very flexible molecules you may need to decreaseOPTTOL to, e.g., 3d-5.

My personal opinion is that many default criteria in Gaussian are not very reasonable and were chosen based on compromise between speed and accuracy with emphasis on calculation's speed.

Best regards,
Alex Granovsky


On Mon Feb 18 '08, Tae-Yun Park wrote
-------------------------------------
>Dear PCGAMESS users,

>While Ive been enjoying the performance of PCGAMESS recently, I found another factor that the performance of PCGAMESS could possibly be compared unfairly if one just uses default settings for geometry optimization. The default settings in PCGAMESS are,
>OPTTOL = 0.00010
>RMS Force = OPTTOL/3 = 0.00003333
>RMAX = 0.1
>RMIN = 0.0015

>In the case of G03, the default settings are,
>Max. Force = 0.00045
>RMS Force = 0.00030
>Max. Displacement = 0.00180
>RMS Displacement = 0.00120

>If OPTTOL and Max. Force have got the same meaning, it looks to me that PCGAMESS settings are more severe, which may result in more iterations and calculation time but more accurate geometry. Same for the case of RMS Force setting. I wonder if this is a correct view.

>If indeed the default settings are not really equivalent, I wonder if the PCGAMESS OPTTOL can be set to 0.00045 to make the similar conditions compared to the case of G03. This makes PCGAMESS RMS Force = OPTTOL/3 = 0.00015, which is still smaller than G03. If it is so, what about the RMAX and RMIN? What values could to be used to have the similar conditions to G03 default settings?

>Finally, if indeed the PCGAMESS default settings are more sever compared to those of G03, is there any specific reason for that?

>Any comments would be highly appreciated.

>Many thanks,

>Tae-Yun
>


Message 4df668afbHW-6625-1023+00.htm, number 274, was posted on Wed Feb 20 '08 at 17:03:35
in reply to 5c714e7ePMq-6622-1391+00.htm

Re^8: How to change rsh to ssh?

Alex Granovsky
gran@classic.chem.msu.su


Hi,


if due to some unknown reasons P4_RSHCOMMAND does not work for you,
I can provide you by the customized binaries with ssh used by default.


Best regards,
Alex Granovsky


On Sun Feb 17 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>On Sun Feb 17 '08, Alexei Popov wrote
>-------------------------------------
>>Hi,

>>first, you are using wrong format for procgrp file.
>>The correct format to start, e.g., PC GAMESS on 16 dual-CPU (or just dual core) boxes is:

>>local 1
>>node1 2 /home/avb/715/pcgamess
>>node2 2 /home/avb/715/pcgamess
>>node3 2 /home/avb/715/pcgamess
>>node4 2 /home/avb/715/pcgamess
>>node5 2 /home/avb/715/pcgamess
>>node6 2 /home/avb/715/pcgamess
>>node7 2 /home/avb/715/pcgamess
>>node8 2 /home/avb/715/pcgamess
>>node9 2 /home/avb/715/pcgamess
>>node10 2 /home/avb/715/pcgamess
>>node11 2 /home/avb/715/pcgamess
>>node12 2 /home/avb/715/pcgamess
>>node13 2 /home/avb/715/pcgamess
>>node14 2 /home/avb/715/pcgamess
>>node15 2 /home/avb/715/pcgamess

>>You need to provide the path to executable in the procgrp file!

>>Second, if statically linked binaries do not work for you, try dynamically-linked ones.

>>Finally, you should export your P4_RSHCOMMAND environment variable!

>>Regards,
>>Alexei
>>
>Hi,

>first of all I've ran PCGAMESS on the every nodes in sequentially.

>Next, I've ran PCGAMESS in SMP parallel mode as written in

>http://classic.chem.msu.su/gran/gamess/comm_line.htmlhttp://classic.chem.msu.su/gran/gamess/comm_line.html
>and
>http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-6610-901+00.htmhttp://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-6610-901+00.htm
>without any problems. :)
>Some jobs has been solved successful.

>Next, I've included in .bash_profile the line

>export P4_RSHCOMMAND=ssh

>and reloging in the system.

>As recomended in

>http://classic.chem.msu.su/gran/gamess/linux_mpich.htmlhttp://classic.chem.msu.su/gran/gamess/linux_mpich.html

>in procgrp-file I wrote lines

>local 0
>n02 2 /home/gorbenko/pcgamess715/pcgamess

>and have tried to run PCGAMESS in parallel mode.

>In answer I recived message

>p0_10425: p4_error: create_procgroup: getpwuid failed: 0

>You gave me recomandation to read

>http://classic.chem.msu.su/gran/gamess/introlinux.htmlhttp://classic.chem.msu.su/gran/gamess/introlinux.html

>I read and made

>http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C35c715226PMq-6622-33+00.htmhttp://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C35c715226PMq-6622-33+00.htm

>I'm using procgrp-file like to your one (see above). At running PCGAMESS I recived in answer

>n02: Connection refused

>with line in proccesses list

>5411 pts/5 SN+ 0:00 /usr/bin/rsh n02 -l gorbenko -n /home/gorbenko/pcgamess715/pcgamess n01 56969 \-p4amslave \-p4yourname n02 \-p4rmrank 1

>What is wrong?

>Regards,

>Vitaliy Gorbenko
>
>
>
>
>
>
>
>


Message 5c7106edPMq-6625-1155+00.htm, number 275, was posted on Wed Feb 20 '08 at 19:15:07
in reply to 4df668afbHW-6625-1023+00.htm

Re^9: How to change rsh to ssh?

Vitaliy Gorbenko
vgorbenko@ukr.net


On Wed Feb 20 '08, Alex Granovsky wrote
---------------------------------------
>Hi,
>
>
>if due to some unknown reasons P4_RSHCOMMAND does not work for you,
>I can provide you by the customized binaries with ssh used by default.
>
>
>Best regards,
>Alex Granovsky
>
Thank you, I would be very grateful! I don't know why P4_RSHCOMMAND does not work in this case :(.
How I can obtaine the your binaries with ssh?

With best regards,
Vitaliy Gorbenko


Message 7c39290dcW1-6626-303+00.htm, number 276, was posted on Thu Feb 21 '08 at 05:03:33
in reply to 4df668afbHW-6625-1021+00.htm

Re^2: PCGAMESS default OPT options compared to G03?

Tae-Yun Park
taeyun0000@gmail.com


Dear Alex,

Many thanks for your reply. It answered all of my questions, which is greatly appreciated.

Thanks again,

Tae-Yun


Message 7c39290dcW1-6626-364-00.htm, number 277, was posted on Thu Feb 21 '08 at 06:04:04
Auto input generation tool accounting for symmetry & unique atoms info?

Tae-Yun Park
taeyun0000@gmail.com


Recently I found PCGAMESS calculation speed enhances a lot when geometry optimization is performed with symmetry & unique atom information of molecules. For example when the geometry optimization for ethylene is performed with C1 symmetry, i.e.,

C1
C6.0 0.00000 0.00000 0.66790
H 1.0 0.00000 0.93000 1.22780
H 1.0 0.00000 -0.93000 1.22780
C 6.0 0.00000 0.00000 -0.66790
H 1.0 0.00000 -0.93000 -1.22780
H 1.0 0.00000 0.93000 -1.22780

it took about 13.7 sec in our machine, while when D2h is used, i.e.,

Dnh 2

C 6.0 0.00000 0.00000 0.66790
H 1.0 0.00000 0.93000 1.22780

the calculation was done in 5.4 sec (more than 250% improvement !!).

Thus, input structure based on symmetry information with unique atoms contributes a lot in terms of calculation speed for geometry optimization.

I have software that catches the symmetry of a molecule, but it does not generate the PCGAMESS input structure with unique atoms only. I wonder if anyone of you knows or has got any tool to generate such structure information that enhances the speed of geometry optimization with PCGAMESS.

There might be an input keyword option that PCGAMESS automatically detects this situation, which Im not aware of yet.

Any comments related to this subject will highly be appreciated.

Many thanks in advance,

Tae-Yun


Message c17d4ed5eOC-6627-710+00.htm, number 278, was posted on Fri Feb 22 '08 at 11:49:57
in reply to 7c39290dcW1-6626-364-00.htm

Re: Auto input generation tool accounting for symmetry & unique atoms info?

Solntsev Pasha
solntsev@univ.kiev.ua


Try gabedit or wxmacmolplt. They can determine symmetry of molecules and write unique atoms suitable for PC GAMESS input file. Source codes of the programs are also available.

Message cb4cda38aIJ-6628-383-00.htm, number 279, was posted on Sat Feb 23 '08 at 06:23:04
principal axis/ extended basis function

Allan Abraham
allanbpadama@yahoo.com.ph


hi.. i am running a program on the adhesion of a metal atom to a certain polymer. unfortunately, the output stated that an illegal performance occured and requires extended basis function. besides there is a problem on the order of principal axis.. how could i fix this? thank you so much for any help regarding this matter...

Message 589dda27bQn-6628-784+00.htm, number 280, was posted on Sat Feb 23 '08 at 13:04:06
in reply to cb4cda38aIJ-6628-383-00.htm

Re: principal axis/ extended basis function

Pedro Silva
pedros@ufp.pt



Please provide your input, otherwise it is very difficult to guess what might be going on..

On Sat Feb 23 '08, Allan Abraham wrote
--------------------------------------
>hi.. i am running a program on the adhesion of a metal atom to a certain polymer. unfortunately, the output stated that an illegal performance occured and requires extended basis function. besides there is a problem on the order of principal axis.. how could i fix this? thank you so much for any help regarding this matter...


Message 4df668afbHW-6628-1037+00.htm, number 281, was posted on Sat Feb 23 '08 at 17:17:06
in reply to 7c39290dcW1-6626-364-00.htm

Re: Auto input generation tool accounting for symmetry & unique atoms info?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. You can try to use $contrl coord=cart
2. Take a look at the excellent PC GAMESS tutorial written by Dr. Kalju Kahn (for link, see the "Tutorial" section of the PC GAMESS homepage), where you'll find more information on this topic.
3. You can use ChemCraft to symmetrize structures and create list of symmetry-unique atoms for PC GAMESS input.

Best regards,
Alex


On Thu Feb 21 '08, Tae-Yun Park wrote
-------------------------------------
>Recently I found PCGAMESS calculation speed enhances a lot when geometry optimization is performed with symmetry & unique atom information of molecules. For example when the geometry optimization for ethylene is performed with C1 symmetry, i.e.,

>C1
>C 6.0 0.00000 0.00000 0.66790
>H 1.0 0.00000 0.93000 1.22780
>H 1.0 0.00000 -0.93000 1.22780
>C 6.0 0.00000 0.00000 -0.66790
>H 1.0 0.00000 -0.93000 -1.22780
>H 1.0 0.00000 0.93000 -1.22780

>it took about 13.7 sec in our machine, while when D2h is used, i.e.,

>Dnh 2

>C 6.0 0.00000 0.00000 0.66790
>H 1.0 0.00000 0.93000 1.22780

>the calculation was done in 5.4 sec (more than 250% improvement !!).

>Thus, input structure based on symmetry information with unique atoms contributes a lot in terms of calculation speed for geometry optimization.

>I have software that catches the symmetry of a molecule, but it does not generate the PCGAMESS input structure with unique atoms only. I wonder if anyone of you knows or has got any tool to generate such structure information that enhances the speed of geometry optimization with PCGAMESS.

>There might be an input keyword option that PCGAMESS automatically detects this situation, which Im not aware of yet.

>Any comments related to this subject will highly be appreciated.

>Many thanks in advance,

>Tae-Yun
>


Message 4df668afbHW-6628-1039+00.htm, number 282, was posted on Sat Feb 23 '08 at 17:19:24
in reply to 5c7106edPMq-6625-1155+00.htm

Re^9: How to change rsh to ssh?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

please contact me directly on this topic.

Best regards,
Alex Granovsky


On Wed Feb 20 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>On Wed Feb 20 '08, Alex Granovsky wrote
>---------------------------------------
>>Hi,
>>
>>
>>if due to some unknown reasons P4_RSHCOMMAND does not work for you,
>>I can provide you by the customized binaries with ssh used by default.
>>
>>
>>Best regards,
>>Alex Granovsky
>>
>Thank you, I would be very grateful! I don't know why P4_RSHCOMMAND does not work in this case :(.
>How I can obtaine the your binaries with ssh?

>With best regards,
>Vitaliy Gorbenko
>


Message cb4cd351aIJ-6628-1070+00.htm, number 283, was posted on Sat Feb 23 '08 at 17:51:47
in reply to 589dda27bQn-6628-784+00.htm

Re^2: principal axis/ extended basis function

Allan Abraham
allanbpadama@yahoo.com.ph


On Sat Feb 23 '08, Pedro Silva wrote
------------------------------------
>>Please provide your input, otherwise it is very difficult to guess what might be going on..

>On Sat Feb 23 '08, Allan Abraham wrote
>--------------------------------------
>>hi.. i am running a program on the adhesion of a metal atom to a certain polymer. unfortunately, the output stated that an illegal performance occured and requires extended basis function. besides there is a problem on the order of principal axis.. how could i fix this? thank you so much for any help regarding this matter...

This message contains the 2 kb attachment
[ abraham.txt ] optimization


Message 589dda27bQn-6628-1104+00.htm, number 284, was posted on Sat Feb 23 '08 at 18:24:52
in reply to cb4cd351aIJ-6628-1070+00.htm

Re^3: principal axis/ extended basis function

Pedro Silva
pedros@ufp.pt


I found your problem: 6-311G is not defined for Cu. Try en effective core potential, like SKBJ.

Message cb4cdc31aIJ-6629-285+00.htm, number 285, was posted on Sun Feb 24 '08 at 04:45:21
in reply to 589dda27bQn-6628-1104+00.htm

Re^4: principal axis/ extended basis function

Allan Abraham
allanbpadama@yahoo.com.ph


hi sir... thank you so much for ur advice.. unfortunately, i am still encountering error.. i'm really sorry for this...

i want to run this using DFT.. this code (attached) was generated using facio wherein i freezed all the atoms except for copper.. kindly comment and recheck my input code... thanks for ur help....

This message contains the 3 kb attachment
[ copper.txt ] optimization


Message 589dda27bQn-6630-107+00.htm, number 286, was posted on Mon Feb 25 '08 at 01:48:44
in reply to cb4cdc31aIJ-6629-285+00.htm

Re^5: principal axis/ extended basis function

Pedro Silva
pedros@ufp.pt


On Sun Feb 24 '08, Allan Abraham wrote
--------------------------------------
>hi sir... thank you so much for ur advice.. unfortunately, i am still encountering error.. i'm really sorry for this...

>i want to run this using DFT.. this code (attached) was generated using facio wherein i freezed all the atoms except for copper.. kindly comment and recheck my input code... thanks for ur help....


I attach the correct input. It is actually explained in the manual, though ;-)

This message contains the 3 kb attachment
[ corrected input.txt ] Notes: I assumed Mult=2. You should check the ground state spin


Message cb4cd20baIJ-6630-297+00.htm, number 287, was posted on Mon Feb 25 '08 at 04:57:13
in reply to 589dda27bQn-6630-107+00.htm

Re^6: principal axis/ extended basis function

Allan Abraham
allanbpadama@yahoo.com.ph


hi sir....

tnx much for the help.. i'll try to study the manual... tnx a lot. i hope it's ok to ask for your assistance nxt time. best regards.


Message c6b48310SZP-6630-1346-00.htm, number 288, was posted on Mon Feb 25 '08 at 22:28:06
Starting PC-GAMESS with LOW priority

Slawomir Janicki
slawomir.janicki@comcast.net


Hi,

I am experimenting wtih using the start command (under WinXP SP2) to set the priority of PC-GAMESS to Low. My command line is something like this:

start /b /low /wait gamess\pcgamess -i %1.inp -o %1.log

This mostly works as it starts PC-GAMESS with the correct parameters and with priority set to Low , however the program does not respond to ^C. It does respond to ^Break and terminates immediately.

I also tried this:

start "%1" /low /wait gamess\pcgamess -i %1.inp -o %1.log

This opens a new window with the name as in %1, but again, no response to ^C.

Any idea how to restore the ^C response?

SJ


Message d42eebd1bHW-6630-1367+00.htm, number 289, was posted on Mon Feb 25 '08 at 22:47:22
in reply to c6b48310SZP-6630-1346-00.htm

Re: Starting PC-GAMESS with LOW priority

Alex Granovsky
gran@classic.chem.msu.su


Hi,

to reduce priority, just use $system idle=.t. $end in your input.

Best regards,
Alex




On Mon Feb 25 '08, Slawomir Janicki wrote
-----------------------------------------
>Hi,

>I am experimenting wtih using the start command (under WinXP SP2) to set the priority of PC-GAMESS to Low. My command line is something like this:

>start /b /low /wait gamess\pcgamess -i %1.inp -o %1.log

>This mostly works as it starts PC-GAMESS with the correct parameters and with priority set to Low , however the program does not respond to ^C. It does respond to ^Break and terminates immediately.

>I also tried this:

>start "%1" /low /wait gamess\pcgamess -i %1.inp -o %1.log

>This opens a new window with the name as in %1, but again, no response to ^C.

>Any idea how to restore the ^C response?

>SJ


Message 93a30118ZML-6631-963-00.htm, number 290, was posted on Tue Feb 26 '08 at 16:03:08
Troubles with NT-MPICH v 1.5.0

Maurizio Ciofalo
omega@unipa.it


Hi,
last days, within our 64 bit Windows PC GAMESS cluster of machines, I tried to upgrade the 1.5.0 NT-MPICH interface (owing to reportedly "many many bugs have been fixed"), which was released in the last weeks of 2007. However, I got bad results due to a fail of actual communication among machines. Typical symptoms are the undertaking by the main node of all the PC GAMESS processes which should have been distributed among the nodes and the fall of the rcluma service in the main node. However, the MPICH lib subdir is correctly addressed n the PATH system variable, the cluster manager service is active in every machine and Windows SP2 Firewall is open. Anyway, everything worked fine with 1.4.0 release. Thus I had to downgrade to the previous MPICH release, eventually.
Any experience of yours with such a new parallel interface?
Thank you for the attention

Message 7c39290dcW1-6631-1013-00.htm, number 291, was posted on Tue Feb 26 '08 at 16:54:28
PCGAMESS performance to the max!

Tae-Yun Park
taeyun0000@gmail.com


Dear PCGAMESS users,

Although everyone knows the excellent performance of PCGAMESS in terms of calculation speed, it looks to me that it is not difficult to miss some of useful options to enhance the performance even more since there are lots of possible option combinations that users may overlook.

Considering the possibility that I may be one of those users, I would like to list out the options Im currently using at the moment, and make these options to be reviewed by the PCGAMESS experts.

Currently, I am performing geometry optimization and frequency calculation followed by raman calculation using a single AMD quad core CPU(Phenom9500) based machine with 4-core parallelization enabled by mpich under CentOS 5.1. Typical molecule to be calculated consists of 25 heavy atoms, e.g., C, N, O, S, suitably connected with hydrogen.

Here are the options for geometry optimization and frequency calculation:

1. For $CONTRL group, I am using unique coordinate with exact symmetry, and turning on the fastints options:
$CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=OPTIMIZE COORD=UNIQUE ICHARG=0 MULT=1 gencon=.t. fstint=.t. reordr=.t. $END

2. For $SYSTEM group, fastdiag option is turned on with the options for Jacobi diagonalization, fast networking, and buffer overflow:
$SYSTEM MWORDS=30 TIMLIM=600000 kdiag=0 nojac=100 aoints=dist mpisnc=.t. $END

3. For $GUESS group, the followings are used to speed up Huckel guess:
$GUESS kdiag=0 guess=huckel $END

4. The basis set used for the calculation is:
$BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END

5. For $STATPT group, the options are:
$STATPT hssend=.t. NSTEP=50 $END

6. For $SCF group, and $INTGRL group, QFMM is turned on:
$SCF dirscf=.t. $END
$INTGRL schwrz=.t. $END

7. For $FMM group, I am using full direct option with PCGAMESS routines for Coulomb:
$FMM method=fulldrct nearj=2 $END

8. For $SMP and $P2P, the following options are used:
$SMP csmtx=.t. call64=.t. $END
$P2P p2p=.t. dlb=.t. $END

For raman calculation, only run type is changed to RAMAN, deleting the hssend, i.e.,
$CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=RAMAN COORD=UNIQUE
ICHARG=0 MULT=1 gencon=.t. fstint=.t. reordr=.t. $END
$SYSTEM MWORDS=30 TIMLIM=600000
kdiag=0 nojac=100 aoints=dist mpisnc=.t. $END
$GUESS kdiag=0 guess=huckel $END
$BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END
$STATPT NSTEP=50 $END
$SCF dirscf=.t. $END
$INTGRL schwrz=.t. $END
$FMM method=fulldrct nearj=2 $END
$SMP csmtx=.t. call64=.t. $END
$P2P p2p=.t. dlb=.t. $END

These are the combinations of PCGAMESS options I found so far to improve the calculation speed of geometry optimization and frequency calculation followed by raman calculation.

I wonder if there is any option that Im missing or incorrectly assigned to improve the calculation speed of PCGAMESS for my run type.

Any suggestion or comment related to this subject is greatly appreciated.

Many thanks in advance,

Tae-Yun


Message 5395f7047Tv-6633-1207-00.htm, number 292, was posted on Thu Feb 28 '08 at 20:07:40
RI-DFT implementation

Denis Zavelev
denis.zavelev@gmail.com


I hope you've heard about PRIRODA package. The main its feature is using RI-DFT instead of usual DFT that makes the computations of medium and even big systems incomparably faster (though I'm not sure that it gives only better performance and not worse results as the price of performance, but anyway it will greatly help at least to make first approximation better than using MM as semi-empirical methods are not applicable for systems containg metal atoms). But due to some reasons (absence of documentation, very complicated obtaining of binaries, only few basis sets available, no visualisation packages, etc) its use seems to be not the reliable way of investigations.

So I'd like to ask, are they any thoughts to add RI-DFT to next versions of PC Gamess?


Message 4df668afbHW-6633-1433+00.htm, number 293, was posted on Thu Feb 28 '08 at 23:53:16
in reply to 93a30118ZML-6631-963-00.htm

Re: Troubles with NT-MPICH v 1.5.0

Alexei Popov
alexei.a.popov@gmail.com


Hi,

That's really very interesting observation! I'm still using v. 1.4 on my cluster but was about to upgrade. Did you submit the bug report to the mp-mpich mailing list or to NT-MPICH authors? Did you get any replies? Please keep us informed.

Regards,
Alexei




On Tue Feb 26 '08, Maurizio Ciofalo wrote
-----------------------------------------
>Hi,
>last days, within our 64 bit Windows PC GAMESS cluster of machines, I tried to upgrade the 1.5.0 NT-MPICH interface (owing to reportedly "many many bugs have been fixed"), which was released in the last weeks of 2007. However, I got bad results due to a fail of actual communication among machines. Typical symptoms are the undertaking by the main node of all the PC GAMESS processes which should have been distributed among the nodes and the fall of the rcluma service in the main node. However, the MPICH lib subdir is correctly addressed n the PATH system variable, the cluster manager service is active in every machine and Windows SP2 Firewall is open. Anyway, everything worked fine with 1.4.0 release. Thus I had to downgrade to the previous MPICH release, eventually.
>Any experience of yours with such a new parallel interface?
>Thank you for the attention
>


Message 4df668afbHW-6634-8+00.htm, number 294, was posted on Fri Feb 29 '08 at 00:08:11
in reply to 7c39290dcW1-6631-1013-00.htm

Re: PCGAMESS performance to the max!

Alex Granovsky
gran@classic.chem.msu.su


Dear Tae-Yun,

just a couple of suggestions from my side

1. Most of the options you are using are on by default.

2. Most of the default PC GAMESS settings are selected to deliver the best performance

>$CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=OPTIMIZE COORD=UNIQUE ICHARG=0 MULT=1 gencon=.t. fstint=.t. reordr=.t. $END

3. Your $contrl group looks fine to me. I would suggest in addition to use DLC internal coordinates for optimization:

 $contrl nzvar=1 $end
 $zmat dlc=.t. auto=.t. $end

4.
>$SYSTEM MWORDS=30 TIMLIM=600000 kdiag=0 nojac=100 aoints=dist mpisnc=.t. $END

I would suggest to drop mpisnc and aoints options.


5.

>6. For $SCF group, and $INTGRL group, QFMM is turned on:
>$SCF dirscf=.t. $END
>$INTGRL schwrz=.t. $END

Actually, this does not turn QFMM on. To turn it on, you need to set additional flag in the $intgrl group. It is generally bad idea to use QFMM for small molecules. For larger systems, it is better to try first whether QFMM speeds up things or not for your particular molecules, then to use or not to use QFMM.

6.
>7. For $FMM group, I am using full direct option with PCGAMESS routines for Coulomb:
>$FMM method=fulldrct nearj=2 $END

I would suggest just to use default settings.

Other parts of your input look reasonable to me.

Best regards,
Alex



On Tue Feb 26 '08, Tae-Yun Park wrote
-------------------------------------
>Dear PCGAMESS users,

>Although everyone knows the excellent performance of PCGAMESS in terms of calculation speed, it looks to me that it is not difficult to miss some of useful options to enhance the performance even more since there are lots of possible option combinations that users may overlook.

>Considering the possibility that I may be one of those users, I would like to list out the options Im currently using at the moment, and make these options to be reviewed by the PCGAMESS experts.

>Currently, I am performing geometry optimization and frequency calculation followed by raman calculation using a single AMD quad core CPU(Phenom9500) based machine with 4-core parallelization enabled by mpich under CentOS 5.1. Typical molecule to be calculated consists of 25 heavy atoms, e.g., C, N, O, S, suitably connected with hydrogen.

>Here are the options for geometry optimization and frequency calculation:

>1. For $CONTRL group, I am using unique coordinate with exact symmetry, and turning on the fastints options:
>$CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=OPTIMIZE COORD=UNIQUE ICHARG=0 MULT=1 gencon=.t. fstint=.t. reordr=.t. $END

>2. For $SYSTEM group, fastdiag option is turned on with the options for Jacobi diagonalization, fast networking, and buffer overflow:
>$SYSTEM MWORDS=30 TIMLIM=600000 kdiag=0 nojac=100 aoints=dist mpisnc=.t. $END




>3. For $GUESS group, the followings are used to speed up Huckel guess:
>$GUESS kdiag=0 guess=huckel $END

>4. The basis set used for the calculation is:
>$BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END

>5. For $STATPT group, the options are:
>$STATPT hssend=.t. NSTEP=50 $END

>6. For $SCF group, and $INTGRL group, QFMM is turned on:
>$SCF dirscf=.t. $END
>$INTGRL schwrz=.t. $END

>7. For $FMM group, I am using full direct option with PCGAMESS routines for Coulomb:
>$FMM method=fulldrct nearj=2 $END

>8. For $SMP and $P2P, the following options are used:
>$SMP csmtx=.t. call64=.t. $END
>$P2P p2p=.t. dlb=.t. $END

>For raman calculation, only run type is changed to RAMAN, deleting the hssend, i.e.,
> $CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=RAMAN COORD=UNIQUE
> ICHARG=0 MULT=1 gencon=.t. fstint=.t. reordr=.t. $END
> $SYSTEM MWORDS=30 TIMLIM=600000
> kdiag=0 nojac=100 aoints=dist mpisnc=.t. $END
> $GUESS kdiag=0 guess=huckel $END
> $BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END
> $STATPT NSTEP=50 $END
> $SCF dirscf=.t. $END
> $INTGRL schwrz=.t. $END
> $FMM method=fulldrct nearj=2 $END
> $SMP csmtx=.t. call64=.t. $END
> $P2P p2p=.t. dlb=.t. $END

>These are the combinations of PCGAMESS options I found so far to improve the calculation speed of geometry optimization and frequency calculation followed by raman calculation.

>I wonder if there is any option that Im missing or incorrectly assigned to improve the calculation speed of PCGAMESS for my run type.

>Any suggestion or comment related to this subject is greatly appreciated.

>Many thanks in advance,

>Tae-Yun
>


Message 4df668afbHW-6634-21+00.htm, number 295, was posted on Fri Feb 29 '08 at 00:21:23
in reply to 5395f7047Tv-6633-1207-00.htm

Re: RI-DFT implementation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Sure PRIRODA is certainly not the first and not the only QC package with RI functionality. I would suggest to consider Turbomole for all persons interested in the state of the art implementations of RI-based methods.

We plan to add the RI-based code to the PC GAMESS in the future. However, the main emphasis of current PC GAMESS' DFT code is on hybrid functionals, where one would need to use not only RI-J but RI-K as well. The latter results in performance gain only for relatively large molecules, where one can consider the use of already programmed QFMM options as the alternative. Finally, the MARI (multipole accelerated RI) is attractive option to implement as well.

Best regards,
Alex Granovsky

On Thu Feb 28 '08, Denis Zavelev wrote
--------------------------------------
>I hope you've heard about PRIRODA package. The main its feature is using RI-DFT instead of usual DFT that makes the computations of medium and even big systems incomparably faster (though I'm not sure that it gives only better performance and not worse results as the price of performance, but anyway it will greatly help at least to make first approximation better than using MM as semi-empirical methods are not applicable for systems containg metal atoms). But due to some reasons (absence of documentation, very complicated obtaining of binaries, only few basis sets available, no visualisation packages, etc) its use seems to be not the reliable way of investigations.

>So I'd like to ask, are they any thoughts to add RI-DFT to next versions of PC Gamess?


Message 5395f7047Tv-6634-1155+00.htm, number 296, was posted on Fri Feb 29 '08 at 19:15:18
in reply to 4df668afbHW-6634-8+00.htm

$SMP group

Denis Zavelev
denis.zavelev@gmail.com


>> $SMP csmtx=.t. call64=.t. $END

BTW, can you tell me what do these options mean and are there any other options in $SMP group? I haven't found it in documentation...


Message 4df668afbHW-6635-879+00.htm, number 297, was posted on Sat Mar 1 '08 at 14:38:49
in reply to 93a30118ZML-6631-963-00.htm

Re: Troubles with NT-MPICH v 1.5.0

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it seems some ordinal numbers were changed in NT-MPICH version 1.5, making it incompatible with the mpibind.nt-mpich.dll binding file. I tested NT-MPICH 1.5 on our cluster and found it working fine with mpibind.mpich.nt.dll binding dynamic library (which does not use ordinals but references only symbolic names of exported functions).

Best regards,
Alex






On Tue Feb 26 '08, Maurizio Ciofalo wrote
-----------------------------------------
>Hi,
>last days, within our 64 bit Windows PC GAMESS cluster of machines, I tried to upgrade the 1.5.0 NT-MPICH interface (owing to reportedly "many many bugs have been fixed"), which was released in the last weeks of 2007. However, I got bad results due to a fail of actual communication among machines. Typical symptoms are the undertaking by the main node of all the PC GAMESS processes which should have been distributed among the nodes and the fall of the rcluma service in the main node. However, the MPICH lib subdir is correctly addressed n the PATH system variable, the cluster manager service is active in every machine and Windows SP2 Firewall is open. Anyway, everything worked fine with 1.4.0 release. Thus I had to downgrade to the previous MPICH release, eventually.
>Any experience of yours with such a new parallel interface?
>Thank you for the attention
>


Message 4df668afbHW-6635-1228+00.htm, number 297, was edited on Sat Mar 1 '08 at 20:27:55
and replaces message 4df668afbHW-6635-879+00.htm

Re: Troubles with NT-MPICH v 1.5.0

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it is possible that some ordinal numbers of exported functions may have been changed in NT-MPICH version 1.5, making it incompatible with the mpibind.nt-mpich.dll binding file. I tested NT-MPICH 1.5 on our cluster and found it working fine with mpibind.mpich.nt.dll binding dynamic library (which does not use ordinals but references only symbolic names of exported functions).

Best regards,
Alex






On Tue Feb 26 '08, Maurizio Ciofalo wrote
-----------------------------------------
>Hi,
>last days, within our 64 bit Windows PC GAMESS cluster of machines, I tried to upgrade the 1.5.0 NT-MPICH interface (owing to reportedly "many many bugs have been fixed"), which was released in the last weeks of 2007. However, I got bad results due to a fail of actual communication among machines. Typical symptoms are the undertaking by the main node of all the PC GAMESS processes which should have been distributed among the nodes and the fall of the rcluma service in the main node. However, the MPICH lib subdir is correctly addressed n the PATH system variable, the cluster manager service is active in every machine and Windows SP2 Firewall is open. Anyway, everything worked fine with 1.4.0 release. Thus I had to downgrade to the previous MPICH release, eventually.
>Any experience of yours with such a new parallel interface?
>Thank you for the attention
>

[ This message was edited on Sat Mar 1 '08 by the author ]


Message 4df668afbHW-6635-1233+00.htm, number 298, was posted on Sat Mar 1 '08 at 20:33:07
in reply to 4df668afbHW-6635-1228+00.htm

Re^2: Troubles with NT-MPICH v 1.5.0

Alex Granovsky
gran@classic.chem.msu.su


One additional comment.

I've checked with standard mpibind.nt-mpich.dll binding file as well and found that PC GAMESS works properly with this file and NT-MPICH v. 1.5.0. Thus, the problems you reported are specific to your particular setup.

Regards,
Alex


On Sat Mar 1 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>it is possible that some ordinal numbers of exported functions may have been changed in NT-MPICH version 1.5, making it incompatible with the mpibind.nt-mpich.dll binding file. I tested NT-MPICH 1.5 on our cluster and found it working fine with mpibind.mpich.nt.dll binding dynamic library (which does not use ordinals but references only symbolic names of exported functions).

>Best regards,
>Alex
>
>
>
>
>
>
>On Tue Feb 26 '08, Maurizio Ciofalo wrote
>-----------------------------------------
>>Hi,
>>last days, within our 64 bit Windows PC GAMESS cluster of machines, I tried to upgrade the 1.5.0 NT-MPICH interface (owing to reportedly "many many bugs have been fixed"), which was released in the last weeks of 2007. However, I got bad results due to a fail of actual communication among machines. Typical symptoms are the undertaking by the main node of all the PC GAMESS processes which should have been distributed among the nodes and the fall of the rcluma service in the main node. However, the MPICH lib subdir is correctly addressed n the PATH system variable, the cluster manager service is active in every machine and Windows SP2 Firewall is open. Anyway, everything worked fine with 1.4.0 release. Thus I had to downgrade to the previous MPICH release, eventually.
>>Any experience of yours with such a new parallel interface?
>>Thank you for the attention
>>


Message 4df668afbHW-6635-1238+00.htm, number 299, was posted on Sat Mar 1 '08 at 20:38:22
in reply to 5395f7047Tv-6634-1155+00.htm

Re: $SMP group

Alex Granovsky
gran@classic.chem.msu.su


Hi,

call64 option is documented at:

http://classic.chem.msu.su/gran/gamess/call64.html

Several other options of $smp group are documented at:

http://classic.chem.msu.su/gran/gamess/smp.html

The csmtx options allows use of critical sections instead of slower mutexes by multithreaded code.

Regards,
Alex




On Fri Feb 29 '08, Denis Zavelev wrote
--------------------------------------
>>> $SMP csmtx=.t. call64=.t. $END

>BTW, can you tell me what do these options mean and are there any other options in $SMP group? I haven't found it in documentation...


Message 558cee35O52-6637-544-00.htm, number 300, was posted on Mon Mar 3 '08 at 09:04:26
Restrictions of change of geometry (TRUDGE)

Khimich
khimich@kinet.chem.msu.ru


How to limit the maximal change of bond length and angle for one step of TRUDGE optimization? It was very useful to system which geometry is close to optimum.

Message 7c39290dcW1-6638-467+00.htm, number 301, was posted on Tue Mar 4 '08 at 07:47:22
in reply to 4df668afbHW-6634-8+00.htm

Re^2: Thanks a lot

Tae-Yun Park
taeyun0000@gmail.com


Dear Alex,

Many thanks for your help. It was amazing that the DLC internal coordinate speeds up by a factor of 4 in my case.

Your comments were of great help, of course.

Thanks again,

Tae-Yun


Message 7c39290dcW1-6638-471-00.htm, number 302, was posted on Tue Mar 4 '08 at 07:51:30
Converting to natural internal coordinate

Tae-Yun Park
taeyun0000@gmail.com


Thanks to Dr. Granovskys recent reply, I found that PCGAMESS speeds up the geometry optimization by a factor of 4 in my case when I change the coordinate system from Cartesian to DLC internal. It was striking to me that the performance is enhanced that much by simply putting an additional option for DLC internal coordinate in the input file.

I went to a bit more about the coordinate system, and according to the common GAMESS manual (REFS.doc), natural internal coordinate is even better than DLC internal. A this time, however, I dont have a good idea to get the internal coordinate system, as Im not very used to it yet.

I usually start with MDL mol files, converting it to Cartesian coordinate for PCGAMESS input with openbabel. Now, the mol files need to be converted to natural internal coordinate. Is this possible? If so, do we have any tool or software to do that under linux command line mode?

Any suggestion or comment would be greatly appreciated.

Tae-Yun


Message 7c39290dcW1-6638-480-00.htm, number 303, was posted on Tue Mar 4 '08 at 08:00:56
Nvib=2 for IR intensities?

Tae-Yun Park
taeyun0000@gmail.com


Dear PCGAMESS users,

Has anyone of you experienced the difference between nvib=1 and nvib=2 for IR intensity calculations? This is under testing at the moment with several molecules, but from the common GAMESS manual (REFS.doc) the nvib=2 should be used for IR intensities, while I dont find a big difference yet, except for more calculation time.

I also wonder the nvib=2 option provided more accurate values for frequencies compared to nvib=1. Im unable to confirm it as I dont have experimental frequency values for my molecules.

Any comment on this subject would be greatly appreciated.

Many thanks in advance,

Tae-Yun


Message 4df668afbHW-6638-1068+00.htm, number 304, was posted on Tue Mar 4 '08 at 17:48:27
in reply to 558cee35O52-6637-544-00.htm

Re: Restrictions of change of geometry (TRUDGE)

Alex Granovsky
gran@classic.chem.msu.su


Hi,

unfortunately PC GAMESS does not have any undocumented options related to runtyp=trudge :-)

Regards,
Alex



On Mon Mar 3 '08, Khimich wrote
-------------------------------
>How to limit the maximal change of bond length and angle for one step of TRUDGE optimization? It was very useful to system which geometry is close to optimum.


Message 7c39290d5VG-6639-850-00.htm, number 305, was posted on Wed Mar 5 '08 at 14:10:31
Faster frequency calculations?

Yangsoo Kim
hit8run@gmail.com


Dear PCGAMESS users,

It looks to me that the frequency calculation at B3LYP/6-31G* runs with OPTIMIZE & hessend = true can only perform with numerical Hessian, not with analytical Hessian under PCGAMESS. I wonder if this is true. When I use the following keywords,

$CONTRL SCFTYP=RHF DFTTYP=B3LYP RUNTYP=OPTIMIZE COORD=CART $END
$BASIS NGAUSS=6 GBASIS=N31 NDFUNC=1 $END
$STATPT hssend=.t. NSTEP=50 $END
$FORCE method=analytic $END

I got error at the starting point of Hessian calculations. I wonder if this error means the analytical Hessian cant be performed or if Im using a bad keyword combination.

The reason for considering this issue is that generally numerical Hessian calculation seems more time consuming than analytical Hessian in frequency calculations. Is this true for frequency calculations with PCGAMESS?

Finally, I wonder if there is any other possibility to speed up the frequency calculation with PCGAMESS. I tried to perform runtype=HESSIAN separately with initial Hessian (hess=read) obtained from lower level of calculation (DFT/STO-3G), but it seems that the standalone runtype=HESSIAN calculation does not read the initial Hessian. Im not sure if this approach helps, and if it is possible to read initial Hessian to the job, run runtype=HESSIAN.

Any suggestion would be greatly appreciated.

Thanks a lot in advance.

Yang Soo


Message 589dda27bQn-6639-1251+00.htm, number 306, was posted on Wed Mar 5 '08 at 20:51:27
in reply to 7c39290d5VG-6639-850-00.htm

Re: Faster frequency calculations?

Pedro Silva
pedros@ufp.pt


Re: your 1st question, only numerical frequencies are available for DFT,
Re: your 2nd request
For runtyp=hessian jobs, you can input a previous hessian with
$force rdhess=.t. $end

It's all in the manual ;-)

Pedro S.


Message 8231aacfErv-6640-173-00.htm, number 307, was posted on Thu Mar 6 '08 at 02:53:27
How to assign separate HDDs

Damodaran Krishnan
damodak@pitt.edu


Dear PCGAMESS users,

How to assign a separate HDD as a scratch to each instance of the parallel PC GAMESS process ?

thanks
Damodaran


Message 7c39290d5VG-6640-341+00.htm, number 308, was posted on Thu Mar 6 '08 at 05:40:51
in reply to 589dda27bQn-6639-1251+00.htm

Re^2: Faster frequency calculations?

Yangsoo Kim
hit8run@gmail.com


Dear Dr. Silva:

>Re: your 1st question, only numerical frequencies are available for DFT,
Thanks. It clears out various wondering points for me.


>Re: your 2nd request
>For runtyp=hessian jobs, you can input a previous hessian with
> $force rdhess=.t. $end
I did try this option. In this case the frequencies are calculated directly from the input hessian without updating it, so that the calculation is infinitively fast but the result frequency values are not in high quality.

What I actually wanted was to do the runtype=hessian, read the initial hessian, and then update the hessian to do the normal frequency calculations like hessend=true run, hoping that separating this frequency calculation step from (runtype=optimize, hessend=true) may provide a chance to enhance the calculation speed by using analytical hessian or any other means. Of course, this is just a thought from my side, which may be useless to think of.

Thanks again for your comments,

Yang Soo


Message c317208dbQn-6640-919+00.htm, number 309, was posted on Thu Mar 6 '08 at 15:19:24
in reply to 7c39290d5VG-6640-341+00.htm

Re^3: Faster frequency calculations?

Pedro Silva
pedros@ufp.pt



>>Re: your 2nd request
>>For runtyp=hessian jobs, you can input a previous hessian with
>> $force rdhess=.t. $end
>I did try this option. In this case the frequencies are calculated directly from the input hessian without updating it, so that the calculation is infinitively fast but the result frequency values are not in high quality.



I'm sorry, but I forgot to add sommthing: reading a previous hessian with $force rdhess=.t. $end , is useful if you want to compute thermochemical parameters with different scaling factors and/or temperatures. But no new hessian is computed, so that the output frequencies are those from the previous hessian.

Pedro S


Message 4df668afbHW-6640-1017+00.htm, number 310, was posted on Thu Mar 6 '08 at 16:57:41
in reply to 8231aacfErv-6640-173-00.htm

Re: How to assign separate HDDs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. Using old-style command line syntax:

It's easy, just put the list of working directories at the end of your PC GAMESS command line. Do not ues -t option in this case! E.g.,

pcgamess -r -f -p -i /home/gran/test.inp -o /home/gran/test.out -ex /home/gran/pcg /tmp1/pcg /tmp2/pcg ...

Here /tmp1/pcg, /tmp2/pcg, etc... are the working directories for the PC GAMESS instances. They can reside on different HDDs if needed.

Note that all /tmp1/pcg, /tmp2/pcg, etc... should exist prior to the PC GAMESS execution!

The similar syntax works under Windows.


2. Using -t command line option.
This is a little bit tricky, but also possible. Assume -t option is as follows: -t /tmp/pcgtest

This means that PC GAMESS will create /tmp/pcgtest.0, /tmp/pcgtest.1, etc... folders during parallel execution. Thus, before launching PC GAMESS, you just need to create the symbolic link (junk point under Windows) linking /tmp/pcgtest.1 to point to, say, /tmp2/scratch, which resides on different HDD.

Hope this helps.

Best regards,
Alex Granovsky


On Thu Mar 6 '08, Damodaran Krishnan wrote
------------------------------------------
>Dear PCGAMESS users,

>How to assign a separate HDD as a scratch to each instance of the parallel PC GAMESS process ?

>thanks
>Damodaran


Message 4df668afbHW-6640-1026+00.htm, number 311, was posted on Thu Mar 6 '08 at 17:06:28
in reply to 7c39290d5VG-6640-341+00.htm

Re^3: Faster frequency calculations?

Alex Granovsky
gran@classic.chem.msu.su


Dear Yangsoo Kim,

>What I actually wanted was to do the runtype=hessian, read the initial hessian, and then update the hessian to do the normal frequency calculations like hessend=true run, hoping that separating this frequency calculation step from (runtype=optimize, hessend=true) may provide a chance to enhance the calculation speed by using analytical hessian or any other means. Of course, this is just a thought from my side, which may be useless to think of.

It is impossible to speedup calculation of hessian using hessian of the lower theory level as a "starting guess". What is indeed possible in many cases is to use low-level theory hessian as the starting guess for geometry optimization at the higher level of theory. E.g., one can use hessians computed at semiempirical level of theory (as they are extremely cheap) to speedup location of equilibrium geometry at RHF or DFT level of theory.

Best regards,
Alex Granovsky


Message 4df668afbHW-6640-1031+00.htm, number 312, was posted on Thu Mar 6 '08 at 17:11:06
in reply to 7c39290dcW1-6638-471-00.htm

Re: Converting to natural internal coordinate

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it's not trivial to construct natural internal coordinates manually. As far as I remember, I have the program which automates this process somewhere in my archives - it was donated by its authors many years ago to the PC GAMESS project but still was not incorporated into PC GAMESS or otherwise used. I'll try to locate it and produce Windows and Linux executables. I do hope this would be of common interest to the PC GAMESS community.

Best regards,
Alex Granovsky


On Tue Mar 4 '08, Tae-Yun Park wrote
------------------------------------
>Thanks to Dr. Granovskys recent reply, I found that PCGAMESS speeds up the geometry optimization by a factor of 4 in my case when I change the coordinate system from Cartesian to DLC internal. It was striking to me that the performance is enhanced that much by simply putting an additional option for DLC internal coordinate in the input file.

>I went to a bit more about the coordinate system, and according to the common GAMESS manual (REFS.doc), natural internal coordinate is even better than DLC internal. A this time, however, I dont have a good idea to get the internal coordinate system, as Im not very used to it yet.

>I usually start with MDL mol files, converting it to Cartesian coordinate for PCGAMESS input with openbabel. Now, the mol files need to be converted to natural internal coordinate. Is this possible? If so, do we have any tool or software to do that under linux command line mode?

>Any suggestion or comment would be greatly appreciated.

>Tae-Yun
>


Message 4df668afbHW-6640-1043+00.htm, number 313, was posted on Thu Mar 6 '08 at 17:23:45
in reply to 7c39290dcW1-6638-480-00.htm

Re: Nvib=2 for IR intensities?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

The nvib=2 frequencies and intensities are indeed somewhat more accurate than nvib=1 frequencies and intensities. In particular, six "zero" frequencies are typically much better when using nvib=2 and this means that other frequencies have much less contributions of rotational and translational contaminants. The latter can be important for non-rigid molecules.

However, in many cases it is sufficient to perform just nvib=1 calculations. In this case, it may be helpful to increase the overall precision of calculations to some degree (e.g., changing icut, nconv, etc...) and to decrease vibsiz to e.g. 0.005.

Note you can easily perform nvib=1 calculations and then simply continue them as nvib=2 job by changing nvib to 2 in your input, adding $vib group from IRCDATA file, and provide converged starting SCF vectors of vib0 point. On the other hand, it is also easy to model nvib=1 situation using outputs of nvib=2 calculations - once again by changing nvib, and providing $vib and $vec groups. You should be careful however not to change vibsiz during all these manipulations!

Best regards,
Alex Granovsky






On Tue Mar 4 '08, Tae-Yun Park wrote
------------------------------------
>Dear PCGAMESS users,

>Has anyone of you experienced the difference between nvib=1 and nvib=2 for IR intensity calculations? This is under testing at the moment with several molecules, but from the common GAMESS manual (REFS.doc) the nvib=2 should be used for IR intensities, while I dont find a big difference yet, except for more calculation time.

>I also wonder the nvib=2 option provided more accurate values for frequencies compared to nvib=1. Im unable to confirm it as I dont have experimental frequency values for my molecules.

>Any comment on this subject would be greatly appreciated.

>Many thanks in advance,

>Tae-Yun
>


Message 81fc185cZ6e-6640-1343-00.htm, number 314, was posted on Thu Mar 6 '08 at 22:22:44
DFT calculation differences

Joseph Flora
flora-joseph@sc.edu


Folks,

I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

Thanks.

Joe


Message 8f84c543aGH-6641-348+00.htm, number 315, was posted on Fri Mar 7 '08 at 05:48:23
in reply to 81fc185cZ6e-6640-1343-00.htm

Re: DFT calculation differences

Roman Zubatyuk
zubatyuk@gmail.com


Hi Joe,

Even using the same grid and functional the exact energies will slightly differ because of different grid purning schemes (not almost sure that this is the only meaningful difference between PC GAMESS and G03). Although, most probably you will see virtually no difference in relative energies, gradients, force constants, etc.

Regards,
Roman Zubatyuk

On Thu Mar 6 '08, Joseph Flora wrote
------------------------------------
>Folks,

>I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

>Thanks.

>Joe


Message 7c39290dcW1-6641-369+00.htm, number 316, was posted on Fri Mar 7 '08 at 06:09:09
in reply to 4df668afbHW-6640-1031+00.htm

Re^2: Converting to natural internal coordinate

Tae-Yun Park
taeyun0000@gmail.com


Dear Alex,

Many thanks for your reply. Considering the importance of calculation speed that everyone may be interested in, having the tool for generating natural internal coordinate would be of great importance. At least this is clearly the case to me.

Therefore, it would be greatly appreciated if you could possibly make it available to PCGAMESS users including myself.

Always feel lots of thanks to you for your efforts related to PCGAMESS development and management.

Sincerely,



Tae-Yun


On Thu Mar 6 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>it's not trivial to construct natural internal coordinates manually. As far as I remember, I have the program which automates this process somewhere in my archives - it was donated by its authors many years ago to the PC GAMESS project but still was not incorporated into PC GAMESS or otherwise used. I'll try to locate it and produce Windows and Linux executables. I do hope this would be of common interest to the PC GAMESS community.

>Best regards,
>Alex Granovsky


Message 4df668afbHW-6641-759+00.htm, number 317, was posted on Fri Mar 7 '08 at 12:39:14
in reply to 81fc185cZ6e-6640-1343-00.htm

Re: DFT calculation differences

Alex Granovsky
gran@classic.chem.msu.su


Hi Joe,

Taking into account that you are getting such a large energy differences, I think there are some errors in your input files for PC GAMESS and/or Gaussian. Could you please attach them to your message for our inspection?

Best regards,
Alex



On Thu Mar 6 '08, Joseph Flora wrote
------------------------------------
>Folks,

>I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

>Thanks.

>Joe


Message d42eff99bHW-6641-1104+00.htm, number 318, was posted on Fri Mar 7 '08 at 18:24:41
in reply to 4df668afbHW-6641-759+00.htm

Re^2: DFT calculation differences

Alex Granovsky
gran@classic.chem.msu.su


Hi again,

another thing which should be checked is whether you got the same electronic states/SCF solutions with both programs.

Regards,
Alex


On Fri Mar 7 '08, Alex Granovsky wrote
--------------------------------------
>Hi Joe,

>Taking into account that you are getting such a large energy differences, I think there are some errors in your input files for PC GAMESS and/or Gaussian. Could you please attach them to your message for our inspection?

>Best regards,
>Alex
>
>
>
>On Thu Mar 6 '08, Joseph Flora wrote
>------------------------------------
>>Folks,

>>I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

>>Thanks.

>>Joe


Message 81fc185cZ6e-6641-1333+00.htm, number 319, was posted on Fri Mar 7 '08 at 22:14:55
in reply to d42eff99bHW-6641-1104+00.htm

Re^3: DFT calculation differences

Joseph Flora
flora-joseph@sc.edu


There are slight differences in the eigen values for the occupied orbitals, and different numbers for the unoccupied orbitals. Attached in 7zip format are the input files and output files. Thanks!

Joe


On Fri Mar 7 '08, Alex Granovsky wrote
--------------------------------------
>Hi again,

>another thing which should be checked is whether you got the same electronic states/SCF solutions with both programs.

>Regards,
>Alex
>
>
>On Fri Mar 7 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi Joe,

>>Taking into account that you are getting such a large energy differences, I think there are some errors in your input files for PC GAMESS and/or Gaussian. Could you please attach them to your message for our inspection?

>>Best regards,
>>Alex
>>
>>
>>
>>On Thu Mar 6 '08, Joseph Flora wrote
>>------------------------------------
>>>Folks,

>>>I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

>>>Thanks.

>>>Joe

This message contains the 14 kb attachment
[ pcg001.7z ] DFT-comparison-pcg-g03w


Message 4df668afbHW-6641-1378+00.htm, number 320, was posted on Fri Mar 7 '08 at 22:58:23
in reply to 81fc185cZ6e-6641-1333+00.htm

Re^4: DFT calculation differences

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I examined your files and it turns out that your problem is related with the fact that Gaussian uses spherical basis functions while PC GAMESS uses Cartesian functions by default. By performing PC GAMESS run with d5 option set to .t. I was able to almost exactly reproduce your Gaussian's numbers.

Best regards,
Alex


On Fri Mar 7 '08, Joseph Flora wrote
------------------------------------
>There are slight differences in the eigen values for the occupied orbitals, and different numbers for the unoccupied orbitals. Attached in 7zip format are the input files and output files. Thanks!

>Joe
>
>
>On Fri Mar 7 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi again,

>>another thing which should be checked is whether you got the same electronic states/SCF solutions with both programs.

>>Regards,
>>Alex
>>
>>
>>On Fri Mar 7 '08, Alex Granovsky wrote
>>--------------------------------------
>>>Hi Joe,

>>>Taking into account that you are getting such a large energy differences, I think there are some errors in your input files for PC GAMESS and/or Gaussian. Could you please attach them to your message for our inspection?

>>>Best regards,
>>>Alex
>>>
>>>
>>>
>>>On Thu Mar 6 '08, Joseph Flora wrote
>>>------------------------------------
>>>>Folks,

>>>>I compared PCGamess and G03W results of a small system with a transition metal and got disturbing results with default settings. I'm using DFT/B3LYP1, SBKJC for the transition metal, and 6-31G(d) for the rest of the atoms. Basis sets were explicitly specified and are identical. I get binding energy differences of 14-15 kcal/mole between PCG and G03W. I've changed nrad in PCG to be equivalent to that of G03W (and vice versa), and no appreciable change in results. LMAX is the same for both defaults. It appears that the main difference is caused by the transition metal. If I compare just the energies (some have said this is not valid because of differences in code - but binding energies should be similar), I get a 2.6 kcal/mol difference for the transition metal alone, 0.6 kcal/mole for the system without the metal, and 16.6 kcal/mol for the system with the metal. Any thoughts on why such large differences and how to reconcile the two?

>>>>Thanks.

>>>>Joe


Message 81fc185cZ6e-6643-1041+00.htm, number 321, was posted on Sun Mar 9 '08 at 17:21:13
in reply to 4df668afbHW-6641-1378+00.htm

Re^5: DFT calculation differences

Joseph Flora
flora-joseph@sc.edu


Thanks! Binding energy differences are now less than 0.1 kcal/mole.

Joe


Message 53e162eb1oc-6643-1146+00.htm, number 322, was posted on Sun Mar 9 '08 at 19:06:13
in reply to 4df668afbHW-6640-1017+00.htm

Re^2: How to assign separate HDDs

Pasquale
morvillo@portici.enea.it


On Thu Mar 6 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>1. Using old-style command line syntax:

>It's easy, just put the list of working directories at the end of your PC GAMESS command line. Do not ues -t option in this case! E.g.,

>pcgamess -r -f -p -i /home/gran/test.inp -o /home/gran/test.out -ex /home/gran/pcg /tmp1/pcg /tmp2/pcg ...

>Here /tmp1/pcg, /tmp2/pcg, etc... are the working directories for the PC GAMESS instances. They can reside on different HDDs if needed.

In this case, will the AOINTS files be splitted?


Message 53e162eb1oc-6643-1148+00.htm, number 322, was edited on Sun Mar 9 '08 at 19:07:59
and replaces message 53e162eb1oc-6643-1146+00.htm

Re^2: How to assign separate HDDs

Pasquale
morvillo@portici.enea.it


On Thu Mar 6 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>1. Using old-style command line syntax:

>It's easy, just put the list of working directories at the end of your PC GAMESS command line. Do not ues -t option in this case! E.g.,

>pcgamess -r -f -p -i /home/gran/test.inp -o /home/gran/test.out -ex /home/gran/pcg /tmp1/pcg /tmp2/pcg ...

>Here /tmp1/pcg, /tmp2/pcg, etc... are the working directories for the PC GAMESS instances. They can reside on different HDDs if needed.

In this case, will the AOINTS files be splitted?

I am running some DFT calculations but the hard disk space is not enought for AOINTS file. I have also another hard disk. There is any chance to use the space of both disks?

[ This message was edited on Sun Mar 9 '08 by the author ]


Message d4c0ee0abHW-6643-1283+00.htm, number 323, was posted on Sun Mar 9 '08 at 21:23:06
in reply to 53e162eb1oc-6643-1148+00.htm

Re^3: How to assign separate HDDs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes, in this case aoints will be sptitted between several disks if running PC GAMESS in parallel. You can also try to create software or hardware RAID volume to overcome disk space limitation, or just switch to direct calculations.

Best regards,
Alex Granovsky


On Sun Mar 9 '08, Pasquale wrote
--------------------------------
>On Thu Mar 6 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>1. Using old-style command line syntax:

>>It's easy, just put the list of working directories at the end of your PC GAMESS command line. Do not ues -t option in this case! E.g.,

>>pcgamess -r -f -p -i /home/gran/test.inp -o /home/gran/test.out -ex /home/gran/pcg /tmp1/pcg /tmp2/pcg ...

>>Here /tmp1/pcg, /tmp2/pcg, etc... are the working directories for the PC GAMESS instances. They can reside on different HDDs if needed.

>In this case, will the AOINTS files be splitted?

>I am running some DFT calculations but the hard disk space is not enought for AOINTS file. I have also another hard disk. There is any chance to use the space of both disks?


Message 53e0c0e01oc-6643-1375+00.htm, number 324, was posted on Sun Mar 9 '08 at 23:28:03
in reply to d4c0ee0abHW-6643-1283+00.htm

Re^4: How to assign separate HDDs

Pasquale
morvillo@portici.enea.it


Hi,

>yes, in this case aoints will be sptitted between several disks if running PC GAMESS in parallel. You can also try to create software or hardware RAID volume to overcome disk space limitation, or just switch to direct calculations.



By the way, I am using the parallel version of pcgamess (pcgamess as downloaded) and I tried to use 2 directories but aoints is stored only in one! I used the sintax you suggested. I have an Intel dual core 2 processor. From task manager, it seem that only one processor is used (CPU time at least 50%, also for direct runs). Do you have some hints to give me?


Message d4c0ee0abHW-6644-13+00.htm, number 323, was edited on Mon Mar 10 '08 at 00:13:15
and replaces message d4c0ee0abHW-6643-1283+00.htm

Re^3: How to assign separate HDDs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes, in this case aoints will be split between several disks if running PC GAMESS in parallel. You can also try to create software or hardware RAID volume to overcome disk space limitation, or just switch to direct calculations.

Best regards,
Alex Granovsky


On Sun Mar 9 '08, Pasquale wrote
--------------------------------
>On Thu Mar 6 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>1. Using old-style command line syntax:

>>It's easy, just put the list of working directories at the end of your PC GAMESS command line. Do not ues -t option in this case! E.g.,

>>pcgamess -r -f -p -i /home/gran/test.inp -o /home/gran/test.out -ex /home/gran/pcg /tmp1/pcg /tmp2/pcg ...

>>Here /tmp1/pcg, /tmp2/pcg, etc... are the working directories for the PC GAMESS instances. They can reside on different HDDs if needed.

>In this case, will the AOINTS files be splitted?

>I am running some DFT calculations but the hard disk space is not enought for AOINTS file. I have also another hard disk. There is any chance to use the space of both disks?

[ This message was edited on Mon Mar 10 '08 by the author ]


Message d4c0ee0abHW-6644-22+00.htm, number 325, was posted on Mon Mar 10 '08 at 00:22:02
in reply to 53e0c0e01oc-6643-1375+00.htm

Re^5: How to assign separate HDDs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Linux mpich-linked version:

just create the proper procgrp file and add the following to the end of the command line options:

 -p4pg /home/pasquale/pcg.pg

assuming the procgrp file is called pcg.pg and is in /home/pasquale directory

Windows version:

0. Read readme.bindings, readme.commandline, and readme.parallel.windows.nt-mpich-smp

1. Copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder, overwriting the default version of the mpibind.dll.

2. Add -np 2 or whatsoever you want to the end of the PC GAMESS command line

Best regards,
Alex Granovsky


On Sun Mar 9 '08, Pasquale wrote
--------------------------------
>Hi,

>>yes, in this case aoints will be sptitted between several disks if running PC GAMESS in parallel. You can also try to create software or hardware RAID volume to overcome disk space limitation, or just switch to direct calculations.
>
>
>
>By the way, I am using the parallel version of pcgamess (pcgamess as downloaded) and I tried to use 2 directories but aoints is stored only in one! I used the sintax you suggested. I have an Intel dual core 2 processor. From task manager, it seem that only one processor is used (CPU time at least 50%, also for direct runs). Do you have some hints to give me?


Message 53e0911a1oc-6644-1371+00.htm, number 326, was posted on Mon Mar 10 '08 at 22:51:47
in reply to d4c0ee0abHW-6644-22+00.htm

Re^6: How to assign separate HDDs

Pasquale
morvillo@portici.enea.it


Hi,

>Windows version:

>0. Read readme.bindings, readme.commandline, and readme.parallel.windows.nt-mpich-smp

>1. Copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder, overwriting the default version of the mpibind.dll.

>2. Add -np 2 or whatsoever you want to the end of the PC GAMESS command line


Now it works well, as expected! Many thanks.


Message 4f0da159ZML-6648-1224+00.htm, number 327, was posted on Fri Mar 14 '08 at 20:23:44
in reply to 4df668afbHW-6640-1043+00.htm

Re^2: Nvib=2 for IR intensities?

Maurizio Ciofalo
omega@unipa.it


>However, in many cases it is sufficient to perform just nvib=1 calculations.

Is among them the case where one needs the hessian of a starting geometry for hunting a TS through the RUNTYP=SADPOINT key? It would be great to save half of the CPU time!


Message 4df668afbHW-6648-1329+00.htm, number 328, was posted on Fri Mar 14 '08 at 22:09:25
in reply to 7c39290dcW1-6641-369+00.htm

Re^3: Converting to natural internal coordinate

Alex Granovsky
gran@classic.chem.msu.su


Hi,

after some minor changes, I was able to compile INTC program for both Windows and Linux. You can find the password-protected rar archive in the "Downloads" section. Please contact me directly to get the password to unrar it.

Hope this helps.

Best regards,
Alex

On Fri Mar 7 '08, Tae-Yun Park wrote
------------------------------------
>Dear Alex,

>Many thanks for your reply. Considering the importance of calculation speed that everyone may be interested in, having the tool for generating natural internal coordinate would be of great importance. At least this is clearly the case to me.

>Therefore, it would be greatly appreciated if you could possibly make it available to PCGAMESS users including myself.

>Always feel lots of thanks to you for your efforts related to PCGAMESS development and management.

>Sincerely,
>
>
>
>Tae-Yun
>
>
>On Thu Mar 6 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>it's not trivial to construct natural internal coordinates manually. As far as I remember, I have the program which automates this process somewhere in my archives - it was donated by its authors many years ago to the PC GAMESS project but still was not incorporated into PC GAMESS or otherwise used. I'll try to locate it and produce Windows and Linux executables. I do hope this would be of common interest to the PC GAMESS community.

>>Best regards,
>>Alex Granovsky


Message d42efec9bHW-6653-38+00.htm, number 329, was posted on Wed Mar 19 '08 at 00:37:34
in reply to 4f0da159ZML-6648-1224+00.htm

Re^3: Nvib=2 for IR intensities?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I think the answer depends on the magnitude of the imaginary frequency of the TS of interest.

Regards,
Alex Granovsky

On Fri Mar 14 '08, Maurizio Ciofalo wrote
-----------------------------------------
>>However, in many cases it is sufficient to perform just nvib=1 calculations.

>Is among them the case where one needs the hessian of a starting geometry for hunting a TS through the RUNTYP=SADPOINT key? It would be great to save half of the CPU time!


Message 8afd7dbbaKC-6654-955-00.htm, number 330, was posted on Thu Mar 20 '08 at 15:55:32
PCM stability

Richard
bonarlaw@liv.ac.uk


Does anyone have suggestions for improving the stability of DFT geometry optimizations using the PCM solvent model? With default parameters, the energy, gradient and bond lengths seem to fluctuate alarmingly (same with the US version).

The molecules are small symmetrical metal complexes (although symmetry is evidently not used with PCM) with a couple of constrained torsions. The relevant keywords I'm using are:

$PCM SOLVENT=WATER $END
$PCMCAV RIN(1)=2.0 RIN(2)=2.0 $END

The PCM options are fairly complex so its difficult to know where to begin...thanks,

Richard


Message 8231aacfErv-6656-45-00.htm, number 331, was posted on Sat Mar 22 '08 at 00:45:36
Parallel DFT calculations

Damodaran Krishnan
damodak@pitt.edu


Dear PCGAMESS users,

I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?


$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
$STATPT NSTEP=9000 $END
$SYSTEM IDLE=.T. MKLNP=2 $END
$P2P P2P=.T. DLB=.T. $END
$SCF DIRSCF=.True. $END
$GUESS GUESS=HUCKEL $END
$BASIS GBASIS=N21 NGAUSS=3 $END
$DATA


thanks
Damodaran


Message 53e14b901oc-6657-690+00.htm, number 332, was posted on Sun Mar 23 '08 at 11:29:46
in reply to 8231aacfErv-6656-45-00.htm

Re: Parallel DFT calculations

Pasquale
morvillo@portici.enea.it


Hi,
I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

Regards

On Sat Mar 22 '08, Damodaran Krishnan wrote
-------------------------------------------
>Dear PCGAMESS users,

>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>
>
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
> $STATPT NSTEP=9000 $END
> $SYSTEM IDLE=.T. MKLNP=2 $END
> $P2P P2P=.T. DLB=.T. $END
> $SCF DIRSCF=.True. $END
> $GUESS GUESS=HUCKEL $END
> $BASIS GBASIS=N21 NGAUSS=3 $END
> $DATA
>
>
>thanks
>Damodaran


Message 8231aacfErv-6657-1280+00.htm, number 333, was posted on Sun Mar 23 '08 at 21:20:38
in reply to 53e14b901oc-6657-690+00.htm

Re^2: Parallel DFT calculations

Damodaran Krishnan
damodak@pitt.edu


Yes, I am using the Windows version and I did copy the dll file. I should have mentioned that I am able to run other calculations in parallel except DFT.

thanks
Damodaran


On Sun Mar 23 '08, Pasquale wrote
---------------------------------
>Hi,
>I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

>Regards

>On Sat Mar 22 '08, Damodaran Krishnan wrote
>-------------------------------------------
>>Dear PCGAMESS users,

>>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>>
>>
>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
>> $STATPT NSTEP=9000 $END
>> $SYSTEM IDLE=.T. MKLNP=2 $END
>> $P2P P2P=.T. DLB=.T. $END
>> $SCF DIRSCF=.True. $END
>> $GUESS GUESS=HUCKEL $END
>> $BASIS GBASIS=N21 NGAUSS=3 $END
>> $DATA
>>
>>
>>thanks
>>Damodaran


Message d42effa7bHW-6657-1419+00.htm, number 334, was posted on Sun Mar 23 '08 at 23:39:31
in reply to 8231aacfErv-6657-1280+00.htm

Re^3: Parallel DFT calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

did you assign two different working directories to PC GAMESS instances?

Note, you do not need mklnp=2 for DFT jobs.

Best regards,
Alex Granovsky


On Sun Mar 23 '08, Damodaran Krishnan wrote
-------------------------------------------
>Yes, I am using the Windows version and I did copy the dll file. I should have mentioned that I am able to run other calculations in parallel except DFT.

>thanks
>Damodaran
>
>
>On Sun Mar 23 '08, Pasquale wrote
>---------------------------------
>>Hi,
>>I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

>>Regards

>>On Sat Mar 22 '08, Damodaran Krishnan wrote
>>-------------------------------------------
>>>Dear PCGAMESS users,

>>>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>>>
>>>
>>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
>>> $STATPT NSTEP=9000 $END
>>> $SYSTEM IDLE=.T. MKLNP=2 $END
>>> $P2P P2P=.T. DLB=.T. $END
>>> $SCF DIRSCF=.True. $END
>>> $GUESS GUESS=HUCKEL $END
>>> $BASIS GBASIS=N21 NGAUSS=3 $END
>>> $DATA
>>>
>>>
>>>thanks
>>>Damodaran


Message 8231aacfErv-6658-164+00.htm, number 335, was posted on Mon Mar 24 '08 at 02:43:56
in reply to d42effa7bHW-6657-1419+00.htm

Re^4: Parallel DFT calculations

Damodaran Krishnan
damodak@pitt.edu


Hi Alex,

I did not assign two different working directories for the two PC GAMESS instances. I could run regular HF in parallel without assigning different working directories. Is it required to assign for DFT calculations ?

thanks
Damodaran

On Sun Mar 23 '08, Alex Granovsky wrote
---------------------------------------
>Hi,

>did you assign two different working directories to PC GAMESS instances?

>Note, you do not need mklnp=2 for DFT jobs.

>Best regards,
>Alex Granovsky
>
>
>On Sun Mar 23 '08, Damodaran Krishnan wrote
>-------------------------------------------
>>Yes, I am using the Windows version and I did copy the dll file. I should have mentioned that I am able to run other calculations in parallel except DFT.

>>thanks
>>Damodaran
>>
>>
>>On Sun Mar 23 '08, Pasquale wrote
>>---------------------------------
>>>Hi,
>>>I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

>>>Regards

>>>On Sat Mar 22 '08, Damodaran Krishnan wrote
>>>-------------------------------------------
>>>>Dear PCGAMESS users,

>>>>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>>>>
>>>>
>>>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
>>>> $STATPT NSTEP=9000 $END
>>>> $SYSTEM IDLE=.T. MKLNP=2 $END
>>>> $P2P P2P=.T. DLB=.T. $END
>>>> $SCF DIRSCF=.True. $END
>>>> $GUESS GUESS=HUCKEL $END
>>>> $BASIS GBASIS=N21 NGAUSS=3 $END
>>>> $DATA
>>>>
>>>>
>>>>thanks
>>>>Damodaran


Message d4c0f196bHW-6658-775+00.htm, number 336, was posted on Mon Mar 24 '08 at 12:54:46
in reply to 8231aacfErv-6658-164+00.htm

Re^5: Parallel DFT calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this is required by almost all types of calculations, with some RHF cases (depending on your exact options) being the exception. Generally, you should consider assignment of different working directories to different PC GAMESS instances as the mandatory requirement for successful parallel run.

Best regards,
Alex


On Mon Mar 24 '08, Damodaran Krishnan wrote
-------------------------------------------
>Hi Alex,

>I did not assign two different working directories for the two PC GAMESS instances. I could run regular HF in parallel without assigning different working directories. Is it required to assign for DFT calculations ?

>thanks
>Damodaran

>On Sun Mar 23 '08, Alex Granovsky wrote
>---------------------------------------
>>Hi,

>>did you assign two different working directories to PC GAMESS instances?

>>Note, you do not need mklnp=2 for DFT jobs.

>>Best regards,
>>Alex Granovsky
>>
>>
>>On Sun Mar 23 '08, Damodaran Krishnan wrote
>>-------------------------------------------
>>>Yes, I am using the Windows version and I did copy the dll file. I should have mentioned that I am able to run other calculations in parallel except DFT.

>>>thanks
>>>Damodaran
>>>
>>>
>>>On Sun Mar 23 '08, Pasquale wrote
>>>---------------------------------
>>>>Hi,
>>>>I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

>>>>Regards

>>>>On Sat Mar 22 '08, Damodaran Krishnan wrote
>>>>-------------------------------------------
>>>>>Dear PCGAMESS users,

>>>>>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>>>>>
>>>>>
>>>>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
>>>>> $STATPT NSTEP=9000 $END
>>>>> $SYSTEM IDLE=.T. MKLNP=2 $END
>>>>> $P2P P2P=.T. DLB=.T. $END
>>>>> $SCF DIRSCF=.True. $END
>>>>> $GUESS GUESS=HUCKEL $END
>>>>> $BASIS GBASIS=N21 NGAUSS=3 $END
>>>>> $DATA
>>>>>
>>>>>
>>>>>thanks
>>>>>Damodaran


Message 8231aacfErv-6658-1351+00.htm, number 337, was posted on Mon Mar 24 '08 at 22:30:55
in reply to d4c0f196bHW-6658-775+00.htm

Re^6: Parallel DFT calculations

Damodaran Krishnan
damodak@pitt.edu


Thanks very much. Parallel DFT calculation works when I assign different working directories to different PC GAMESS instances.

thanks
Damodaran

On Mon Mar 24 '08, Alex Granovsky wrote
---------------------------------------
>Hi,

>this is required by almost all types of calculations, with some RHF cases (depending on your exact options) being the exception. Generally, you should consider assignment of different working directories to different PC GAMESS instances as the mandatory requirement for successful parallel run.

>Best regards,
>Alex
>
>
>On Mon Mar 24 '08, Damodaran Krishnan wrote
>-------------------------------------------
>>Hi Alex,

>>I did not assign two different working directories for the two PC GAMESS instances. I could run regular HF in parallel without assigning different working directories. Is it required to assign for DFT calculations ?

>>thanks
>>Damodaran

>>On Sun Mar 23 '08, Alex Granovsky wrote
>>---------------------------------------
>>>Hi,

>>>did you assign two different working directories to PC GAMESS instances?

>>>Note, you do not need mklnp=2 for DFT jobs.

>>>Best regards,
>>>Alex Granovsky
>>>
>>>
>>>On Sun Mar 23 '08, Damodaran Krishnan wrote
>>>-------------------------------------------
>>>>Yes, I am using the Windows version and I did copy the dll file. I should have mentioned that I am able to run other calculations in parallel except DFT.

>>>>thanks
>>>>Damodaran
>>>>
>>>>
>>>>On Sun Mar 23 '08, Pasquale wrote
>>>>---------------------------------
>>>>>Hi,
>>>>>I suppose you are using the Windows version of PCGAMESS. In this case, did you copy mpibind.nt-mpich-smp.dll into the mpibind.dll in the PC GAMESS installation folder (to enable parallel mode)?

>>>>>Regards

>>>>>On Sat Mar 22 '08, Damodaran Krishnan wrote
>>>>>-------------------------------------------
>>>>>>Dear PCGAMESS users,

>>>>>>I am not able to run DFT calculations in parallel. When I run it without "-np 2" it runs fine. But with "-np 2" the run gets killed immediately. Here is my input, is there anything wrong?
>>>>>>
>>>>>>
>>>>>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP UNITS=ANGS COORD=ZMT $END
>>>>>> $STATPT NSTEP=9000 $END
>>>>>> $SYSTEM IDLE=.T. MKLNP=2 $END
>>>>>> $P2P P2P=.T. DLB=.T. $END
>>>>>> $SCF DIRSCF=.True. $END
>>>>>> $GUESS GUESS=HUCKEL $END
>>>>>> $BASIS GBASIS=N21 NGAUSS=3 $END
>>>>>> $DATA
>>>>>>
>>>>>>
>>>>>>thanks
>>>>>>Damodaran


Message d4c0f196bHW-6659-943+00.htm, number 338, was posted on Tue Mar 25 '08 at 15:42:52
in reply to 8afd7dbbaKC-6654-955-00.htm

Re: PCM stability

Alex Granovsky
gran@classic.chem.msu.su


Hi Richard,

could you please attach your complete input file to this thread?

Best regards,
Alex


On Thu Mar 20 '08, Richard wrote
--------------------------------
>Does anyone have suggestions for improving the stability of DFT geometry optimizations using the PCM solvent model? With default parameters, the energy, gradient and bond lengths seem to fluctuate alarmingly (same with the US version).

>The molecules are small symmetrical metal complexes (although symmetry is evidently not used with PCM) with a couple of constrained torsions. The relevant keywords I'm using are:
>
>$PCM SOLVENT=WATER $END
>$PCMCAV RIN(1)=2.0 RIN(2)=2.0 $END

>The PCM options are fairly complex so its difficult to know where to begin...thanks,

>Richard


Message 8afd7dbbaKC-6659-1158+00.htm, number 339, was posted on Tue Mar 25 '08 at 19:20:18
in reply to d4c0f196bHW-6659-943+00.htm

Re^2: PCM stability

Richard
bonarlaw@liv.ac.uk


Small example attached - it seems to spend a long time in PCManalyze when run (compressed output to be sent separately).

Richard


On Tue Mar 25 '08, Alex Granovsky wrote
---------------------------------------
>Hi Richard,

>could you please attach your complete input file to this thread?

>Best regards,
>Alex
>
>
>On Thu Mar 20 '08, Richard wrote
>--------------------------------
>>Does anyone have suggestions for improving the stability of DFT geometry optimizations using the PCM solvent model? With default parameters, the energy, gradient and bond lengths seem to fluctuate alarmingly (same with the US version).

>>The molecules are small symmetrical metal complexes (although symmetry is evidently not used with PCM) with a couple of constrained torsions. The relevant keywords I'm using are:
>>
>>$PCM SOLVENT=WATER $END
>>$PCMCAV RIN(1)=2.0 RIN(2)=2.0 $END

>>The PCM options are fairly complex so its difficult to know where to begin...thanks,

>>Richard

This message contains the 23 kb attachment
[ pcm_inp.txt ]


Message 935b012dcvZ-6660-964-00.htm, number 340, was posted on Wed Mar 26 '08 at 16:03:58
What basis set in pcgamess is equivalent to gaussian TZVP?

Zoran Matovic
zmatovic@kg.ac.yu


Dear PCGAMESS users,

just a small practical quest: If I wish to use TZVP basis set (I used earlier in gaussian) whats the equal bs in pcgamess. I serched for tzvp in http://www.emsl.pnl.gov/forms/basisform.html library but there is only tzvp for H-Ar range without transition metals (G03W include TZVP for H-Kr including firsst seria of transition metals). I need this bs for Ni and Cu. Please help as I'm in hurry.

thanks in advance

have a nice day people
Zoran


Message 4df668afbHW-6661-837+00.htm, number 341, was posted on Thu Mar 27 '08 at 13:57:27
in reply to 935b012dcvZ-6660-964-00.htm

Re: What basis set in pcgamess is equivalent to gaussian TZVP?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there is no built-in basis set in PC GAMESS directly corresponding to TZVP. Create and use external basis set file.

Best regards,
Alex


On Wed Mar 26 '08, Zoran Matovic wrote
--------------------------------------
>Dear PCGAMESS users,

>just a small practical quest: If I wish to use TZVP basis set (I used earlier in gaussian) whats the equal bs in pcgamess. I serched for tzvp in http://www.emsl.pnl.gov/forms/basisform.htmlhttp://www.emsl.pnl.gov/forms/basisform.html library but there is only tzvp for H-Ar range without transition metals (G03W include TZVP for H-Kr including firsst seria of transition metals). I need this bs for Ni and Cu. Please help as I'm in hurry.
>
>thanks in advance

>have a nice day people
>Zoran
>


Message 8231aacfErv-6662-74-00.htm, number 342, was posted on Fri Mar 28 '08 at 01:14:34
Optimization of excited states

Damodaran Krishnan
damodak@pitt.edu


Dear users,

Does the current version of PCGAMESS support optimization of excited state geometries? I get the following error message,

*** ERROR *** TDHF & TDDFT GRADIENTS ARE NOT YET PROGRAMMED

thanks
Damodaran


Message 435ddbd2Rew-6662-1338-00.htm, number 343, was posted on Fri Mar 28 '08 at 22:18:17
Running on 2 dual core Xeon (Proc 5060) cpus

Randy Miles
rdmsgl@gmail.com


I am unable to utilize the full processing capacity of this
computer. It has 2 physical cpus on one board with shared memory.
Each physical cpu has 4 logical processors. I am using
mpibind.wmpi-1.3.dll, and the following procgrp file:

local 0
localhost 1 C:\PCGAMESS\PG2\pcgamess.exe
localhost 1 C:\PCGAMESS\PG3\pcgamess.exe
localhost 1 C:\PCGAMESS\PG4\pcgamess.exe
localhost 1 C:\PCGAMESS\PG5\pcgamess.exe
localhost 1 C:\PCGAMESS\PG6\pcgamess.exe
localhost 1 C:\PCGAMESS\PG7\pcgamess.exe
localhost 1 C:\PCGAMESS\PG8\pcgamess.exe

In this case, two pcgamess instances are initiated, but they both
hand indefinitely, producing no output. If I try the following
procgrp file

local 7

I get 8 pcgamess instances initiated and valid output, but they are
confined to 4 of the 8 processors, utilizing only half of my
resources.

I recognize that for efficiency I should place each of the working
directories on separate disks, and I plan to do that in the future.

Is there something else I should be doing differently?

Thanks-


Message 4df668afbHW-6663-1306+00.htm, number 344, was posted on Sat Mar 29 '08 at 21:47:16
in reply to 435ddbd2Rew-6662-1338-00.htm

Re: Running on 2 dual core Xeon (Proc 5060) cpus

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your system has four processor cores. Moreover, as your processors support HTT, and it seems it is enabled, you have eight logical cores available. However, this does not mean that running eight PC GAMESS processes in parallel would be a good idea.

For discussion of some HTT-related topics, look at:

http://classic.chem.msu.su/gran/gamess/smp.html

In brief, I would recommend you to use only four logical processors for calculations, and thus the following procgrp file:

local 3

In this case, PC GAMESS will bind most of its computation threads to the specific logical instances of the HTT-enabled cores for better efficiency, and will decide dynamically whether to use other logical cores during its execution, or not at all.

For the alternative options, please see the link above.


Another point is that the WMPI 1.3 distribution is outdated and buggy. The recommended MPI implementation for Windows PC GAMESS version is NT-MPICH or NT-MPICH-SMP. There are already several recent threads on this forum discussing how to run PC GAMESS in parallel using these MPI implementations.


Best regards,
Alex Granovsky




On Fri Mar 28 '08, Randy Miles wrote
------------------------------------
>I am unable to utilize the full processing capacity of this
>computer. It has 2 physical cpus on one board with shared memory.
>Each physical cpu has 4 logical processors. I am using
>mpibind.wmpi-1.3.dll, and the following procgrp file:

>local 0
>localhost 1 C:\PCGAMESS\PG2\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG3\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG4\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG5\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG6\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG7\pcgamess.exe
>localhost 1 C:\PCGAMESS\PG8\pcgamess.exe

>In this case, two pcgamess instances are initiated, but they both
>hand indefinitely, producing no output. If I try the following
>procgrp file

>local 7

>I get 8 pcgamess instances initiated and valid output, but they are
>confined to 4 of the 8 processors, utilizing only half of my
>resources.

>I recognize that for efficiency I should place each of the working
>directories on separate disks, and I plan to do that in the future.

>Is there something else I should be doing differently?

>Thanks-
>


Message 4df668afbHW-6663-1315+00.htm, number 345, was posted on Sat Mar 29 '08 at 21:54:56
in reply to 8231aacfErv-6662-74-00.htm

Re: Optimization of excited states

Alex Granovsky
gran@classic.chem.msu.su


Hi,


at present, TDHF and TDDFT gradients are not supported. This work has been started recently and one may expect the updated binaries within several months.

This does not mean you cannot deal with the excited states using PC GAMESS. First, gradients are available for CI jobs, including efficient large-scale code for CIS gradients. Second, you can use MCSCF and state averaged MCSCF to optimize geometries of the excited states. Finally, if your molecule is small you can use any available method of the excited states energy calculation with runtyp=TRUDGE.

Best regards,
Alex Granovsky


On Fri Mar 28 '08, Damodaran Krishnan wrote
-------------------------------------------
>Dear users,

>Does the current version of PCGAMESS support optimization of excited state geometries? I get the following error message,

>*** ERROR *** TDHF & TDDFT GRADIENTS ARE NOT YET PROGRAMMED

>thanks
>Damodaran


Message 53b8cce0Juh-6665-17-00.htm, number 346, was posted on Mon Mar 31 '08 at 00:19:22
TDDFT results accuracy

Enrico
e.chiavazza@gmail.com


I would do same uv/vis simulation on cyanine dye but, pc-gamess TDDFT B3LYP/6-31G(d,p) response of a optimized structure is about 100 nm blue-shifted respect sperimental (solvent used for the misure were CH3OH, so I've simulated it with PCM) results and about 50 nm respect a gaussian 03w calculation result of the same level, so I don't know if there is some other parameter to adjust.
I have already read the available tdtft example on pcgamess readmes and tutorial but i don't find any an explanation of this differencies.

I attach my input so you can control if i have do some mistake.

This message contains the 5 kb attachment
[ dMi_MS_TDDFT.inp ] TDDFT input file


Message 4df668afbHW-6665-945+00.htm, number 347, was posted on Mon Mar 31 '08 at 15:45:20
in reply to 53b8cce0Juh-6665-17-00.htm

Re: TDDFT results accuracy

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there is no mistakes in your input file.

I'm just curious why do you expect getting the experimental quality values from the relatively low-level theoretical calculations?

Best regards,
Alex Granovsky


On Mon Mar 31 '08, Enrico wrote
-------------------------------
>I would do same uv/vis simulation on cyanine dye but, pc-gamess TDDFT B3LYP/6-31G(d,p) response of a optimized structure is about 100 nm blue-shifted respect sperimental (solvent used for the misure were CH3OH, so I've simulated it with PCM) results and about 50 nm respect a gaussian 03w calculation result of the same level, so I don't know if there is some other parameter to adjust.
>I have already read the available tdtft example on pcgamess readmes and tutorial but i don't find any an explanation of this differencies.

>I attach my input so you can control if i have do some mistake.


Message 53b8ca37Juh-6666-182-00.htm, number 348, was posted on Tue Apr 1 '08 at 03:01:44
Fsf: fatal error

Enrico
e.chiavazza@gmail.com


Hi,
I'm trying for follow your suggest so I've launched my tddft calculation changing functional and basis set, PBE0/6-311+G**, but when i try to add the $PCM option it crash with "FSF: Fatal error no. 0x00000020 in sub SEQOPN on unit 27" on both check and run runtyp.

without $PCM option it works well on 4 node SMP-mode

Thanks in advance!


Message 53b8ca37Juh-6666-956+00.htm, number 349, was posted on Tue Apr 1 '08 at 15:56:05
in reply to 53b8ca37Juh-6666-182-00.htm

Re: Fsf: fatal error-> solved

Enrico
e.chiavazza@gmail.com


May be i've understood what's wrong:

After several trial i've changed the pc-gamess dir from c:/documents and setting/enrico/....ecc to c:/gms without any space in the path now all work well also the pcm

bye


On Tue Apr 1 '08, Enrico wrote
------------------------------
>Hi,
>I'm trying for follow your suggest so I've launched my tddft calculation changing functional and basis set, PBE0/6-311+G**, but when i try to add the $PCM option it crash with "FSF: Fatal error no. 0x00000020 in sub SEQOPN on unit 27" on both check and run runtyp.

>without $PCM option it works well on 4 node SMP-mode

>Thanks in advance!


Message 4df668afbHW-6667-1290+00.htm, number 350, was posted on Wed Apr 2 '08 at 21:31:28
in reply to 8afd7dbbaKC-6659-1158+00.htm

Re^3: PCM stability

Alex Granovsky
gran@classic.chem.msu.su


Dear Richard,

your sample is indeed the difficult case to handle. If you look on the attached picture with the overlapping Van-der-Waals spheres of your compound, you'll find that small variations of geometry can cause dramatic changes in the topology of spheres and hence the tesselation process is unstable. Indeed, the character of overlap between spheres of central part of your molecule and DMFA ligands varies dramatically wrt. to small changes of geometry. Unfortunately I cannot provide you by the easy solution of this problem. However, several words of advise are as follows:

1. Always optimize your system in gas phase first before doing PCM
2. Changing omega, fro and ret may help generate more smooth and stable surface.

Best regards,
Alex



On Tue Mar 25 '08, Richard wrote
--------------------------------
>Small example attached - it seems to spend a long time in PCManalyze when run (compressed output to be sent separately).

>Richard
>
>
>On Tue Mar 25 '08, Alex Granovsky wrote
>---------------------------------------
>>Hi Richard,

>>could you please attach your complete input file to this thread?

>>Best regards,
>>Alex
>>
>>
>>On Thu Mar 20 '08, Richard wrote
>>--------------------------------
>>>Does anyone have suggestions for improving the stability of DFT geometry optimizations using the PCM solvent model? With default parameters, the energy, gradient and bond lengths seem to fluctuate alarmingly (same with the US version).

>>>The molecules are small symmetrical metal complexes (although symmetry is evidently not used with PCM) with a couple of constrained torsions. The relevant keywords I'm using are:
>>>
>>>$PCM SOLVENT=WATER $END
>>>$PCMCAV RIN(1)=2.0 RIN(2)=2.0 $END

>>>The PCM options are fairly complex so its difficult to know where to begin...thanks,

>>>Richard

This message contains the 73 kb attachment
[ pcm_spheres.jpg ] PCM spheres


Message 4df668afbHW-6667-1295+00.htm, number 351, was posted on Wed Apr 2 '08 at 21:35:36
in reply to 53b8ca37Juh-6666-956+00.htm

Re^2: Fsf: fatal error-> solved

Alex Granovsky
gran@classic.chem.msu.su


Hi,

if you like to use spaces in arguments passed to PC GAMESS you should use e.g.,

  -i "C:\Program files\PCGAMESS\test.inp"

but not

  -i C:\Program files\PCGAMESS\test.inp

and similarly for other command line options.

Regards,
Alex Granovsky


On Tue Apr 1 '08, Enrico wrote
------------------------------
>May be i've understood what's wrong:

>After several trial i've changed the pc-gamess dir from c:/documents and setting/enrico/....ecc to c:/gms without any space in the path now all work well also the pcm

>bye
>
>
>On Tue Apr 1 '08, Enrico wrote
>------------------------------
>>Hi,
>>I'm trying for follow your suggest so I've launched my tddft calculation changing functional and basis set, PBE0/6-311+G**, but when i try to add the $PCM option it crash with "FSF: Fatal error no. 0x00000020 in sub SEQOPN on unit 27" on both check and run runtyp.

>>without $PCM option it works well on 4 node SMP-mode

>>Thanks in advance!


Message 53b8ca37Juh-6668-107-00.htm, number 352, was posted on Thu Apr 3 '08 at 01:46:59
Pc-Gamess mpich version on multicore

Enrico
e.chiavazza@gmail.com


Hi,
new day, new problem!
I would run the linux version of pcgamess mpich linked on a multicore pc (quadcore), but after the succesfully installation of MPICH 1.2.7p1 with the configure option --with device=ch_shmem, I have seen that mpirun -np=4 /path/pcgamess -r -f -o file.out results in an error in mpi device calling of the program (ch_p4).
But if i try to configure mpich with ch_p4 device and i launch a calculation, it result in:

mpirun -np 4 /home/enrico/bin/gms/pcgamess -r -f -i td_dMi_MeO_6311bis.inp -o td_dMi_MeO_6311bis.log -t /tmp/scrach
p0_11059: p4_error: Path to program is invalid while starting /home/enrico/bin/gms/pcgamess with rsh on Q6600: -1
p4_error: latest msg from perror: No such file or directory

Under windows with NT-mpi dll all four core works well but i don't have enough memory (winxp sp2 read only 3.25gb of 4 installed) so i need to work under opensuse 10.3 default kernel.

thanks



Message d4c0ee0b2GM-6668-897+00.htm, number 353, was posted on Thu Apr 3 '08 at 14:58:07
in reply to 53b8ca37Juh-6668-107-00.htm

Re: Pc-Gamess mpich version on multicore

Igor Polyakov
polyakoviv@gmail.com


Hello, Enrico!
On Thu Apr 3 '08, Enrico wrote
------------------------------
>Hi,
>new day, new problem!
>I would run the linux version of pcgamess mpich linked on a multicore pc (quadcore), but after the succesfully installation of MPICH 1.2.7p1 with the configure option --with device=ch_shmem, I have seen that mpirun -np=4 /path/pcgamess -r -f -o file.out results in an error in mpi device calling of the program (ch_p4).

I suppose u have to use device ch_p4 for pc gamess/mpich.

>But if i try to configure mpich with ch_p4 device and i launch a calculation, it result in:
>mpirun -np 4 /home/enrico/bin/gms/pcgamess -r -f -i td_dMi_MeO_6311bis.inp -o td_dMi_MeO_6311bis.log -t /tmp/scrach
>p0_11059: p4_error: Path to program is invalid while starting /home/enrico/bin/gms/pcgamess with rsh on Q6600: -1
> p4_error: latest msg from perror: No such file or directory

Check your procgrp or arch file for your machine.
If you do not have rsh daemon running, i suppose u have ssh daemon running (the most popular case), this way u should set envirinmental P4_RSHCOMMAND equal to SSH, like this:
export P4_RSHCOMMAND=SSH
This was for bash.
Also the thing is for ur case it is better to use static linked pc gamess executable. In this case u do not need to compile and bother urself with mpich and its configuration. To use static link binaries download them, then just create a procgroup file like "procgrp" with the content like "local 3" and in the command line u should specify this file with -p4pg /path/to/the/file/procgrp. Local 3 means 1 master process with id 0 and 3 slave processes. This is exactly the case u need for quad-core smp system.

>Under windows with NT-mpi dll all four core works well but i don't have enough memory (winxp sp2 read only 3.25gb of 4 installed) so i need to work under opensuse 10.3 default kernel.

Actually u do not have to. U can use WinXP 64 bit - it will work ok with 4Gb or win2003sp 32bit (in case u preffer 32 bit) - still works ok with 4Gb.

>thanks
>
>

Wish u good luck!


Message 864742e80zh-6669-189-00.htm, number 354, was posted on Fri Apr 4 '08 at 03:09:49
TDHF Calculation for Pt complex

Stephen Dutz
smdutz@csupomona.edu


Hi all,

I am trying to calculate the hyperpolarizability of a few transition metal (Pt) compounds. To do this I am using the LANL2DZ basis set. After optimizing I perform a tdhf calculation to obtain the hyperpolarizability. When I look at the output of the run, in the hyperpolarizability section it repeatedly prints "MAXIMUM RESPONSE VECTOR ERROR= ...". I've tried searching online for this sort of error, however haven't found much. Could someone please enlighten me as to why this is occurring, and perhaps how to fix it. My input file is below.

Thank you in advance!

Steve

-----------------
$CONTRL RUNTYP=tdhf EXETYP=Run COORD=Unique ICHARG=0 MULT=1 MAXIT=150
ECP=READ NOSYM=1
$END
$DATA
Pt2C10P4H16Cl2
Cnv 2

CARBON 6.0 1.2037363946 0.6952924231 0.0143946952
S 7
1 4233.0000000 0.0012200
2 634.9000000 0.0093420
3 146.1000000 0.0454520
4 42.5000000 0.1546570
5 14.1900000 0.3588660
6 5.1480000 0.4386320
7 1.9670000 0.1459180
S 2
1 5.1480000 -0.1683670
2 0.4962000 1.0600910
S 1
1 0.1533000 1.0000000
P 4
1 18.1600000 0.0185390
2 3.9860000 0.1154360
3 1.1430000 0.3861880
4 0.3594000 0.6401140
P 1
1 0.1146000 1.0000000

CARBON 6.0 0.0000000000 1.4143715030 0.0189893384
S 7
1 4233.0000000 0.0012200
2 634.9000000 0.0093420
3 146.1000000 0.0454520
4 42.5000000 0.1546570
5 14.1900000 0.3588660
6 5.1480000 0.4386320
7 1.9670000 0.1459180
S 2
1 5.1480000 -0.1683670
2 0.4962000 1.0600910
S 1
1 0.1533000 1.0000000
P 4
1 18.1600000 0.0185390
2 3.9860000 0.1154360
3 1.1430000 0.3861880
4 0.3594000 0.6401140
P 1
1 0.1146000 1.0000000

HYDROGEN 1.0 2.1324352343 1.2303859979 0.0005215781
S 3
1 19.2384000 0.0328280
2 2.8987000 0.2312040
3 0.6535000 0.8172260
S 1
1 0.1776000 1.0000000

CARBON 6.0 0.0000000000 2.8563629968 0.0252414625
S 7
1 4233.0000000 0.0012200
2 634.9000000 0.0093420
3 146.1000000 0.0454520
4 42.5000000 0.1546570
5 14.1900000 0.3588660
6 5.1480000 0.4386320
7 1.9670000 0.1459180
S 2
1 5.1480000 -0.1683670
2 0.4962000 1.0600910
S 1
1 0.1533000 1.0000000
P 4
1 18.1600000 0.0185390
2 3.9860000 0.1154360
3 1.1430000 0.3861880
4 0.3594000 0.6401140
P 1
1 0.1146000 1.0000000

CARBON 6.0 0.0000000000 4.0679109839 0.0018380680
S 7
1 4233.0000000 0.0012200
2 634.9000000 0.0093420
3 146.1000000 0.0454520
4 42.5000000 0.1546570
5 14.1900000 0.3588660
6 5.1480000 0.4386320
7 1.9670000 0.1459180
S 2
1 5.1480000 -0.1683670
2 0.4962000 1.0600910
S 1
1 0.1533000 1.0000000
P 4
1 18.1600000 0.0185390
2 3.9860000 0.1154360
3 1.1430000 0.3861880
4 0.3594000 0.6401140
P 1
1 0.1146000 1.0000000

PLATINUM 78.0 0.0000000000 6.0565897780 -0.0119921811
S 3
1 2.5470000 -1.4739175
2 1.6140000 1.9115719
3 0.5167000 0.3922319
S 4
1 2.5470000 1.4388166
2 1.6140000 -2.0911821
3 0.5167000 -1.0921315
4 0.2651000 1.3426596
S 1
1 0.0580000 1.0000000
P 3
1 2.9110000 -0.5247438
2 1.8360000 0.9671884
3 0.5982000 0.5438632
P 2
1 0.6048000 -0.1061438
2 0.0996000 1.0383102
P 1
1 0.0290000 1.0000000
D 2
1 1.2430000 0.5598150
2 0.4271000 0.5511090
D 1
1 0.1370000 1.0000000

PHOSPHORUS 15.0 0.0000000000 5.7981452327 2.3789708735
S 2
1 1.5160000 -0.5862089
2 0.3369000 1.2994376
S 1
1 0.1211000 1.0000000
P 2
1 3.7050000 -0.0691472
2 0.3934000 1.0161988
P 1
1 0.1190000 1.0000000

HYDROGEN 1.0 -1.0938454574 5.0679556741 2.8671755199
S 3
1 19.2384000 0.0328280
2 2.8987000 0.2312040
3 0.6535000 0.8172260
S 1
1 0.1776000 1.0000000

HYDROGEN 1.0 0.0000000000 6.9199956242 3.2328072643
S 3
1 19.2384000 0.0328280
2 2.8987000 0.2312040
3 0.6535000 0.8172260
S 1
1 0.1776000 1.0000000

PHOSPHORUS 15.0 0.0000000000 8.5606005304 -0.1407243865
S 2
1 1.5160000 -0.5862089
2 0.3369000 1.2994376
S 1
1 0.1211000 1.0000000
P 2
1 3.7050000 -0.0691472
2 0.3934000 1.0161988
P 1
1 0.1190000 1.0000000

HYDROGEN 1.0 1.0801971004 9.2742306178 0.4185615336
S 3
1 19.2384000 0.0328280
2 2.8987000 0.2312040
3 0.6535000 0.8172260
S 1
1 0.1776000 1.0000000

HYDROGEN 1.0 0.0000000000 9.0321795967 -1.4614929699
S 3
1 19.2384000 0.0328280
2 2.8987000 0.2312040
3 0.6535000 0.8172260
S 1
1 0.1776000 1.0000000

CHLORINE 17.0 0.0000000000 6.1249434840 -2.4318271227
S 2
1 2.2310000 -0.4900589
2 0.4720000 1.2542684
S 1
1 0.1631000 1.0000000
P 2
1 6.2960000 -0.0635641
2 0.6333000 1.0141355
P 1
1 0.1819000 1.0000000

$END
$ECP
C-ECP NONE
C-ECP NONE
C-ECP NONE
C-ECP NONE
C-ECP NONE
C-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
C-ECP NONE
C-ECP NONE
C-ECP NONE
C-ECP NONE
PT-ECP GEN 60 4
5 ----- g potential -----
-0.1619268 1 728.9394056
-1320.2873852 2 320.6567800
-298.3178135 2 52.8680174
-87.5837065 2 12.0280128
-8.1493274 2 3.5238913
6 ----- s-g potential -----
2.7334218 0 409.4437358
59.7024329 1 274.5419231
891.4589550 2 127.5658570
368.4467656 2 32.9036631
238.0263090 2 5.0593880
-107.0556454 2 4.1506556
4 ----- p-g potential -----
1.8878568 0 466.1728892
76.0138629 1 120.7888259
343.5511116 2 36.4118791
119.4911786 2 5.6985408
5 ----- d-g potential -----
2.9343678 0 249.5650763
59.3306571 1 126.6678585
452.4445194 2 63.1430586
210.4769479 2 17.9059470
58.6254112 2 4.2239373
5 ----- f-g potential -----
3.9534253 0 121.8158799
53.8555182 1 60.8757030
247.4305133 2 31.4767147
127.8187976 2 9.8811751
15.3772046 2 2.7319874
PT-ECP
P-ECP GEN 10 2
5 ----- d potential -----
-10.0000000 1 462.1211423
-79.4864658 2 93.6863701
-28.3668251 2 21.2349094
-9.8577589 2 6.3388415
-1.0163783 2 2.0620684
5 ----- s-d potential -----
3.0000000 0 78.0831823
12.9104154 1 58.9576810
150.0250298 2 36.0571255
71.7083146 2 11.2464453
23.0397012 2 2.6757561
6 ----- p-d potential -----
5.0000000 0 75.1617880
6.3446507 1 57.4544041
198.5585104 2 47.9481748
111.1470820 2 18.4588360
40.3944144 2 5.9414190
6.4483233 2 1.8487507
P-ECP
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
P-ECP
P-ECP
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
H-ECP NONE
CL-ECP GEN 10 2
5 ----- d potential -----
-10.0000000 1 94.8130000
66.2729170 2 165.6440000
-28.9685950 2 30.8317000
-12.8663370 2 10.5841000
-1.7102170 2 3.7704000
5 ----- s-d potential -----
3.0000000 0 128.8391000
12.8528510 1 120.3786000
275.6723980 2 63.5622000
115.6777120 2 18.0695000
35.0606090 2 3.8142000
6 ----- p-d potential -----
5.0000000 0 216.5263000
7.4794860 1 46.5723000
613.0320000 2 147.4685000
280.8006850 2 48.9869000
107.8788240 2 13.2096000
15.3439560 2 3.1831000
CL-ECP
$END
$SCF NCONV=10 FDIFF=.FALSE.
$END
$STATPT NSTEP=300
$END
$SYSTEM TIMLIM=100000 MWORDS=10
$END
$TDHF FREQ(1)=0.04282
$END


Message 4df668afbHW-6669-964+00.htm, number 355, was posted on Fri Apr 4 '08 at 16:03:45
in reply to 864742e80zh-6669-189-00.htm

Re: TDHF Calculation for Pt complex

Alex Granovsky
gran@classic.chem.msu.su


Hi,

This is not the error message at all! The purpose of this message is to control the convergence behavior of CPHF or CPKS procedure. When the error in the responce vector falls below requested threshold, the iterative CPHF solver stops iterate and consider results as converged. At this stage, it is possible to calculate polarizabilities. Thus, you can either safely ignore these messages looking only on the strings of printout containing information on alpha values, or to monitor them estimating how much time you'll need to get the final answer.

Best regards,
Alex


On Fri Apr 4 '08, Stephen Dutz wrote
------------------------------------
>Hi all,

>I am trying to calculate the hyperpolarizability of a few transition metal (Pt) compounds. To do this I am using the LANL2DZ basis set. After optimizing I perform a tdhf calculation to obtain the hyperpolarizability. When I look at the output of the run, in the hyperpolarizability section it repeatedly prints "MAXIMUM RESPONSE VECTOR ERROR= ...". I've tried searching online for this sort of error, however haven't found much. Could someone please enlighten me as to why this is occurring, and perhaps how to fix it. My input file is below.

>Thank you in advance!

>Steve


Message 864742e80zh-6669-1264+00.htm, number 356, was posted on Fri Apr 4 '08 at 21:04:34
in reply to 4df668afbHW-6669-964+00.htm

Re: TDHF Calculation for Pt complex

Stephen Dutz
smdutz@csupomona.edu


Thank you for the clarification Alex!

Sincerely,
Steve

On Fri Apr 4 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>This is not the error message at all! The purpose of this message is to control the convergence behavior of CPHF or CPKS procedure. When the error in the responce vector falls below requested threshold, the iterative CPHF solver stops iterate and consider results as converged. At this stage, it is possible to calculate polarizabilities. Thus, you can either safely ignore these messages looking only on the strings of printout containing information on alpha values, or to monitor them estimating how much time you'll need to get the final answer.

>Best regards,
>Alex
>
>
>On Fri Apr 4 '08, Stephen Dutz wrote
>------------------------------------
>>Hi all,

>>I am trying to calculate the hyperpolarizability of a few transition metal (Pt) compounds. To do this I am using the LANL2DZ basis set. After optimizing I perform a tdhf calculation to obtain the hyperpolarizability. When I look at the output of the run, in the hyperpolarizability section it repeatedly prints "MAXIMUM RESPONSE VECTOR ERROR= ...". I've tried searching online for this sort of error, however haven't found much. Could someone please enlighten me as to why this is occurring, and perhaps how to fix it. My input file is below.

>>Thank you in advance!

>>Steve

>


Message c883fc02S7q-6672-1106-00.htm, number 357, was posted on Mon Apr 7 '08 at 18:26:21
Problens with Hessian Calculation

Marcos Ribeiro
ramribeiro@gmail.com


I tried perform a Hessian Calculation, and the job show me the follow message:

-------------------------------
PARTIAL INTEGRAL TRANSFORMATION
-------------------------------

NUMBER OF CORE MOLECULAR ORBITALS = 0
NUMBER OF OCCUPIED MOLECULAR ORBITALS = 82
TOTAL NUMBER OF MOLECULAR ORBITALS = 492
TOTAL NUMBER OF ATOMIC ORBITALS = 492
THRESHOLD FOR KEEPING TRANSFORMED 2E- INTEGRALS = 1.000E-09
AO INTEGRALS WILL BE READ IN FROM DISK...
# OF WORDS AVAILABLE = 29989374
# OF WORDS NEEDED = 598992598 FOR IN MEMORY TRANSFORMATION
FOR THE SEGMENTED TRANSFORMATION:
MINIMUM= 60789038 WORDS FOR 1 MOLECULAR ORBITAL PER PASS
MAXIMUM= 598992598 WORDS FOR ALL MOLECULAR ORBITALS IN 1 PASS
( 59668776 EXTRA WORDS WOULD INCLUDE AN EXTRA ORBITAL/PASS)

CHOOSING OUT-OF-MEMORY BIN-SORT TRANSFORMATION...
# PASSES = 170
# WORDS USED = 29962901
OPENING FILE DASORT WITH 37023 LOGICAL RECORDS OF 121278 WORDS
DASORT WILL CONTAIN A MAXIMUM OF 37023 PHYSICAL RECORDS OF LENGTH 121280 WORDS

*****************THE Problem*********************
FSF: fatal error no. 0xFFFFFFFB in sub WRITEREC on unit 20


EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 20:16:32 LT 4-APR-2008
30000000 WORDS OF DYNAMIC MEMORY USED
14855 WORDS OF GUARD POOL MEMORY USED

CPU TIME: STEP = 174.95 , TOTAL = 1976.8 S

What's the problem? I already check the INPUT and change the Memory value.

Thank' a lot
Best Regards


Message c883fc02S7q-6672-1106+00.htm, number 357, was edited on Mon Apr 7 '08 at 18:28:40
and replaces message c883fc02S7q-6672-1106-00.htm

Problens with Hessian Calculation

Marcos Ribeiro
ramribeiro@gmail.com


Hi all

I tried perform a Hessian Calculation, and the job show me the follow message:

-------------------------------
PARTIAL INTEGRAL TRANSFORMATION
-------------------------------

NUMBER OF CORE MOLECULAR ORBITALS = 0
NUMBER OF OCCUPIED MOLECULAR ORBITALS = 82
TOTAL NUMBER OF MOLECULAR ORBITALS = 492
TOTAL NUMBER OF ATOMIC ORBITALS = 492
THRESHOLD FOR KEEPING TRANSFORMED 2E- INTEGRALS = 1.000E-09
AO INTEGRALS WILL BE READ IN FROM DISK...
# OF WORDS AVAILABLE = 29989374
# OF WORDS NEEDED = 598992598 FOR IN MEMORY TRANSFORMATION
FOR THE SEGMENTED TRANSFORMATION:
MINIMUM= 60789038 WORDS FOR 1 MOLECULAR ORBITAL PER PASS
MAXIMUM= 598992598 WORDS FOR ALL MOLECULAR ORBITALS IN 1 PASS
( 59668776 EXTRA WORDS WOULD INCLUDE AN EXTRA ORBITAL/PASS)

CHOOSING OUT-OF-MEMORY BIN-SORT TRANSFORMATION...
# PASSES = 170
# WORDS USED = 29962901
OPENING FILE DASORT WITH 37023 LOGICAL RECORDS OF 121278 WORDS
DASORT WILL CONTAIN A MAXIMUM OF 37023 PHYSICAL RECORDS OF LENGTH 121280 WORDS

*****************THE Problem*********************
FSF: fatal error no. 0xFFFFFFFB in sub WRITEREC on unit 20


EXECUTION OF GAMESS TERMINATED ABNORMALLY AT 20:16:32 LT 4-APR-2008
30000000 WORDS OF DYNAMIC MEMORY USED
14855 WORDS OF GUARD POOL MEMORY USED

CPU TIME: STEP = 174.95 , TOTAL = 1976.8 S

What's the problem? I already check the INPUT and change the Memory value.

Thank' a lot
Best Regards

[ This message was edited on Mon Apr 7 '08 by the author ]


Message 4df668afbHW-6673-984+00.htm, number 358, was posted on Tue Apr 8 '08 at 16:24:09
in reply to c883fc02S7q-6672-1106+00.htm

Re: Problens with Hessian Calculation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

This message means that you are out of free disk space.

BIN-SORT transformation is slow and requires lots of disk space but less memory than segmented transformation. I would suggest you to either increase mwords to use segmented transformation (provided you have enough memory) or to bypass integral transformation at all by setting:

  $cphf cphf=ao $end

Best regards,
Alex Granovsky


Message 3e0b5b5fJuh-6675-35+00.htm, number 359, was posted on Thu Apr 10 '08 at 00:34:46
in reply to 4df668afbHW-6673-984+00.htm

Re^2: Problens with Hessian Calculation

Enrico
e.chiavazza@gmail.com


Hi,
I've also a problem with vibrational analysis:
I've run succesfully an optimization (it ends with "***** EQUILIBRIUM GEOMETRY LOCATED ***** statement" but at the end of hessian run of the optimized structure (I've copied the last coordinate set of the log file in pasted it into the $data group of the new input file) I can read the resulting vibrational frequency (all positives) but also the phrase:

*******************************************************
* THIS IS NOT A STATIONARY POINT ON THE MOLECULAR PES *
* THE VIBRATIONAL ANALYSIS IS NOT VALID !!! *
*******************************************************


What it means?
I need to tight the convergence criteria of the geometry search??
$STATPT OPTTOL=1.0E-3 NSTEP=100 METHOD=QA HESS=GUESS $END

Thanks to all

On Tue Apr 8 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>This message means that you are out of free disk space.

>BIN-SORT transformation is slow and requires lots of disk space but less memory than segmented transformation. I would suggest you to either increase mwords to use segmented transformation (provided you have enough memory) or to bypass integral transformation at all by setting:

>

  $cphf cphf=ao $end

>Best regards,
>Alex Granovsky


Message 8afd7fc1aKC-6675-93+00.htm, number 360, was posted on Thu Apr 10 '08 at 01:33:40
in reply to 3e0b5b5fJuh-6675-35+00.htm

Re^3: Problens with Hessian Calculation

Richard
bonarlaw@liv.ac.uk


Yes, OPTTOL=1.0e-3 is much too high, you need 1.0e-4 to 1.0e-6, depending on your molecule. If its not already there adding the line
$FORCE PURIFY=.t. NVIB=2 $END
to your input is a good idea as well.

Richard

On Thu Apr 10 '08, Enrico wrote
-------------------------------
>Hi,
>I've also a problem with vibrational analysis:
>I've run succesfully an optimization (it ends with "***** EQUILIBRIUM GEOMETRY LOCATED ***** statement" but at the end of hessian run of the optimized structure (I've copied the last coordinate set of the log file in pasted it into the $data group of the new input file) I can read the resulting vibrational frequency (all positives) but also the phrase:
>
> *******************************************************
> * THIS IS NOT A STATIONARY POINT ON THE MOLECULAR PES *
> * THE VIBRATIONAL ANALYSIS IS NOT VALID !!! *
> *******************************************************
>
>
>What it means?
>I need to tight the convergence criteria of the geometry search??
>$STATPT OPTTOL=1.0E-3 NSTEP=100 METHOD=QA HESS=GUESS $END

>Thanks to all


Message 8afd7dbbaKC-6691-940+00.htm, number 361, was posted on Sat Apr 26 '08 at 15:41:09
in reply to 4df668afbHW-6667-1290+00.htm

Re^4: PCM stability

Richard
bonarlaw@liv.ac.uk


mail exchange pasted below in case anyone finds the information useful


Hi Richard
I do not know any program allowing display of tessellation details for GAMESS/PC GAMESS.
Frankly, I would be happy to have such a program myself for debugging/PCM related development purposes...

Yes, you should use the 1999 manual which is available at the PC GAMESS homepage.
There is a couple of additional keywords in $pcm group, namely direct=.t./.f. allows you
to select algorithm which uses less disk but more computations. analyz=0 will disable
PCMAnalize feature if required. Almost all of PCM parts are parallelised, the exception
is the construction of the inverse D matrix.

One additional point to know - as internally PCM gradients are only semianalytical, it is usually good idea to loose opttol to e.g. 1d-3 or so.

Best regards,
Alex


>Thanks Alex,
>Do you know of a program that allows one to visualise the details of >the tasselation and PCM surface actually being used by GAMESS (as >opposed to the default ChemCraft vdw radii)?
>
>Should I be going by the the '99 version of the GAMESS manual as far >as PCM in PC-GAMESS is concerned? And are there currently any >limitations in parallelization, as in the current version of >GAMESS-US (MODPAR)
>Richard



On Wed Apr 2 '08, Alex Granovsky wrote
--------------------------------------
>Dear Richard,

>your sample is indeed the difficult case to handle. If you look on the attached picture with the overlapping Van-der-Waals spheres of your compound, you'll find that small variations of geometry can cause dramatic changes in the topology of spheres and hence the tesselation process is unstable. Indeed, the character of overlap between spheres of central part of your molecule and DMFA ligands varies dramatically wrt. to small changes of geometry. Unfortunately I cannot provide you by the easy solution of this problem. However, several words of advise are as follows:

>1. Always optimize your system in gas phase first before doing PCM
>2. Changing omega, fro and ret may help generate more smooth and stable surface.

>Best regards,
>Alex
>
>
>
>On Tue Mar 25 '08, Richard wrote
>--------------------------------
>>Small example attached - it seems to spend a long time in PCManalyze when run (compressed output to be sent separately).

>>Richard
>>
>>
>>On Tue Mar 25 '08, Alex Granovsky wrote
>>---------------------------------------
>>>Hi Richard,

>>>could you please attach your complete input file to this thread?

>>>Best regards,
>>>Alex
>>>
>>>
>>>On Thu Mar 20 '08, Richard wrote
>>>--------------------------------
>>>>Does anyone have suggestions for improving the stability of DFT geometry optimizations using the PCM solvent model? With default parameters, the energy, gradient and bond lengths seem to fluctuate alarmingly (same with the US version).

>>>>The molecules are small symmetrical metal complexes (although symmetry is evidently not used with PCM) with a couple of constrained torsions. The relevant keywords I'm using are:
>>>>
>>>>$PCM SOLVENT=WATER $END
>>>>$PCMCAV RIN(1)=2.0 RIN(2)=2.0 $END

>>>>The PCM options are fairly complex so its difficult to know where to begin...thanks,

>>>>Richard


Message 8afd7dbbaKC-6704-123-00.htm, number 362, was posted on Fri May 9 '08 at 02:03:18
TDDFT/TDA_restarts

Richard
bonarlaw@liv.ac.uk


I don't pretend to understand how these calculations work, but....having run a TDDFT/TDA job with a particular value of NSTATE, should it be possible to rerun the job with a larger value for NSTATE, making use of the already calculated data so the first set of states don't have to be redone i.e. just carry on? It seems the program expects the number of states in the TDVEC group to match NSTATE.

thanks

Richard


Message 52a6dfc2aQp-6704-1251-00.htm, number 363, was posted on Fri May 9 '08 at 20:50:53
pcgamess -t switch and parallel runs

Avishai Ish-Shalom
avishaii@huji.ac.il


hi all.

i've set pcgamess to run in parallel on a 6 node quad core cluster. when runing the command:
~> /opt/pcgamess_dyn/pcgamess -i input -r -f -b /opt/pcgamess/basis_sets/cc-pvdz.lib -ex /opt/pcgamess_dyn -p4pg ~/pgfile /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai /home/scr/avishai
or the equivalent mpirun form, pcgamess runs fine. however, when using the -t switch instead of listing the 6 temp directories, pcgamess just hangs and works forever.

a second related question - if running pcgamess on more than one core of a node, will the -t switch work? the docs say -t directory will be the same for each node, however if a node is running multiple pcgamess sessions using the same directory that could cause problems...

almost forgot, i'm running sles 10sp1 on all nodes, core2 quad cpu's using kernel 2.6.16, mpich 1.2.7 and pcgamess 7.15 dynamically linked to mpich.

regards,
avishai


Message d4c0fb6d2GM-6705-214+00.htm, number 364, was posted on Sat May 10 '08 at 03:33:50
in reply to 52a6dfc2aQp-6704-1251-00.htm

Re: pcgamess -t switch and parallel runs

Igor Polyakov
polyakoviv@gmail.com


hello!

Frankly speaking, just setting -t switch like this -t /scr/avishai in the end should work. I should also point out, that in your commandline i saw something like /home/scr/avishai, but if you use nfs on this cluster and you have /home/scr/ as a common directory you shouldn't use it. For scratch data you should only use local disk of your node. For me works just fine -t /tmp/pcgamess or -t /scr/pcgamess. Using -t switch you will be able to run 4 compute processes on each node to fully utilise your computational power. Of course u need to have appropriate procgrp file.

>almost forgot, i'm running sles 10sp1 on all nodes, core2 quad cpu's using kernel 2.6.16, mpich 1.2.7 and pcgamess 7.15 dynamically linked to mpich.

I thought that dynamically linked version of pcgamess works only when initiated by mpirun (or mpiexec) command in the first place, like mpirun pcgamess , but in your commandline there is simple pcgamess -p4pg which is usually used for static linked versions. But maybe i'm wrong with this one.

>regards,
>avishai
>

Good luck!


Message 4df668afbHW-6706-668+00.htm, number 365, was posted on Sun May 11 '08 at 11:08:29
in reply to d4c0fb6d2GM-6705-214+00.htm

Re^2: pcgamess -t switch and parallel runs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

some minor comments to all involved into discussion:

1. -t switch requires one only one argument, the base of the
temporary file dirs. E.g., if it is: -t /tmp/pcg, each PC GAMESS
instance will create directory /tmp/pcg.#nodenumber and will use
it for scratch files. Thus you'll get directories /tmp/pcg.0, tmp/pcg.1,
etc... This resolves any possible conflicts with scratch files.

2. If you are using the alternative syntax with explicit directories
list, they should exist prior to PC GAMESS execution and they should
be chosen so that to avoid any scratch file conflicts - i.e., they
should all be different on the same node

3. Temporary directories should always reside on the local filesystems.

4. MPICH PC GAMESS executable which is dynamically linked is linked
statically against mpich libraries and dynamically against all other,
system libraries. This allow PC GAMESS to run in parallel or in
sequential mode even without installed MPICH at all. Other PC
GAMESS/dynamically linked versions are linked dynamically with MPI
libraries they are using.

Regards,
Alex Granovsky








Message 4df668afbHW-6706-673+00.htm, number 366, was posted on Sun May 11 '08 at 11:13:27
in reply to 8afd7dbbaKC-6704-123-00.htm

Re: TDDFT/TDA_restarts

Alex Granovsky
gran@classic.chem.msu.su


Dear Richard,

while it is impossible at present (just because this is not
programmed), there should not be any serious problems with
implementation of this feature. At present, old CIS/TD vectors
are only used as a starting guess for the subsequent calculations
at the distorted geometries (e.g., geometry optimization).

Best regards,
Alex



On Fri May 9 '08, Richard wrote
-------------------------------
>I don't pretend to understand how these calculations work, but....having run a TDDFT/TDA job with a particular value of NSTATE, should it be possible to rerun the job with a larger value for NSTATE, making use of the already calculated data so the first set of states don't have to be redone i.e. just carry on? It seems the program expects the number of states in the TDVEC group to match NSTATE.

>thanks

>Richard


Message 3db21250So7-6706-1255-00.htm, number 367, was posted on Sun May 11 '08 at 20:55:20
How can I fix the GRADIENT OUT OF RANGE

ganymede
wuxb07@lzu.cn


Dear Prof. ?3CBR> I am deeply sorry that i have to interrupt you again! ?3CBR> I have a gamess input files about XeF2, which use "Ahlrichs Coulomb Fitting"
Basis Set, In the execution, the DENSITY CONVERGED! but at the end, It appears a error:
?????MAXIMUM GRADIENT = 17.3754560 ?RMS GRADIENT = 8.1908685
0OPTIMIZATION ABORTED.
??- GRADIENT OUT OF RANGE
??- MAXIMUM ALLOWED FORCE (FMAXT) = ?10.000
How can I handle or fix it! thank you very much!
?Best regards!

PS: the input file
$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP MAXIT=200 MULT=1$END
$SYSTEM TIMLIM=525600 MEMORY=10000000 $END
$SCF DIRSCF=.TRUE. $END
$STATPT OPTTOL=0.00001 NSTEP=200 $END
$DATA
Title
DNH 8

Xe ??4.0 ??0.00000 ??0.00000 ??0.00000
S ??4
?? ???????1.35700920E+01 ????0.02573819 ???
?? ???????6.09401700E+00 ????0.21613184 ???
?? ???????2.93348100E+00 ????0.94374761 ???
?? ???????1.50036800E+00 ????1.68889682 ???
S ??1
?? ???????8.06602000E-01 ????0.31316239 ???
S ??1
?? ???????4.50061000E-01 ????0.91731083 ???
S ??1
?? ???????2.56962000E-01 ????0.17208349 ???
S ??1
?? ???????1.47859000E-01 ????0.04586558 ???
P ??1
?? ???????1.21957600E+00 ????1.00000000 ???
P ??1
?? ???????4.86856000E-01 ????1.00000000 ???
P ??1
?? ???????2.02856000E-01 ????1.00000000 ???
D ??2
?? ???????1.46102300E+00 ????0.01707032 ???
?? ???????5.43002000E-01 ????0.10550766 ???
D ??1
?? ???????2.02307000E-01 ????0.00317570 ???
F ??1
?? ???????4.85844740E-01 ????1.00000000 ???
G ??1
?? ???????5.32000000E-01 ????1.00000000 ???

F ??9.0 ??0.00000 ??0.00000 ??1.98769
S ??3
?? ???????1.15670231E+03 ????0.71431186 ???
?? ???????3.45040302E+02 ????1.84554195 ???
?? ???????1.11332058E+02 ????4.71095270 ???
S ??1
?? ???????3.85902340E+01 ????4.21329515 ???
S ??1
?? ???????1.42414660E+01 ????0.11772222 ???
S ??1
?? ???????5.53413200E+00 ????0.15894702 ???
S ??1
?? ???????2.23509600E+00 ????1.49637682 ???
S ??1
?? ???????9.24514000E-01 ????0.68056750 ???
S ??1
?? ???????3.85516000E-01 ????0.11297897 ???
P ??1
?? ???????3.12693419E+00 ????1.00000000 ???
P ??1
?? ???????1.26740199E+00 ????1.00000000 ???
P ??1
?? ???????5.28084163E-01 ????1.00000000 ???
D ??1
?? ???????2.31855700E+01 ????0.04106421 ???
D ??1
?? ???????6.68248300E+00 ????0.13327013 ???
D ??1
?? ???????2.16767700E+00 ????0.12350252 ???
D ??1
?? ???????7.33849000E-01 ????0.03195039 ???
F ??1
?? ???????2.55707762E+00 ????1.00000000 ???

$END


Message 4df668afbHW-6706-1287+00.htm, number 368, was posted on Sun May 11 '08 at 21:27:08
in reply to 3db21250So7-6706-1255-00.htm

Re: How can I fix the GRADIENT OUT OF RANGE

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. You should never use Coulomb (exchange, density, etc...) fitting basis sets (which are used in density fitting aka RI or DF calculations) instead of basis sets which are developed to describe orbitals, not density. PC GAMESS does not use density fitting at moment and thus it does not require DF basis sets. Instead, you should provide "normal" basis set.

2. Use basis sets of correct type and start with a reasonable geometry.

Regards,
Alex Granovsky



On Sun May 11 '08, ganymede wrote
---------------------------------
>Dear Prof. ?3CBR> I am deeply sorry that i have to interrupt you again! ?3CBR> I have a gamess input files about XeF2, which use "Ahlrichs Coulomb Fitting"
>Basis Set, In the execution, the DENSITY CONVERGED! but at the end, It appears a error:
> ????? MAXIMUM GRADIENT = 17.3754560 ? RMS GRADIENT = 8.1908685
>0OPTIMIZATION ABORTED.
> ??- GRADIENT OUT OF RANGE
> ??- MAXIMUM ALLOWED FORCE (FMAXT) = ?10.000
>How can I handle or fix it! thank you very much!
> ?Best regards!

>PS: the input file
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP MAXIT=200 MULT=1$END
> $SYSTEM TIMLIM=525600 MEMORY=10000000 $END
> $SCF DIRSCF=.TRUE. $END
> $STATPT OPTTOL=0.00001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>Xe ??4.0 ??0.00000 ??0.00000 ??0.00000
> S ??4
> ?? ???????1.35700920E+01 ????0.02573819 ???
> ?? ???????6.09401700E+00 ????0.21613184 ???
> ?? ???????2.93348100E+00 ????0.94374761 ???
> ?? ???????1.50036800E+00 ????1.68889682 ???
> S ??1
> ?? ???????8.06602000E-01 ????0.31316239 ???
> S ??1
> ?? ???????4.50061000E-01 ????0.91731083 ???
> S ??1
> ?? ???????2.56962000E-01 ????0.17208349 ???
> S ??1
> ?? ???????1.47859000E-01 ????0.04586558 ???
> P ??1
> ?? ???????1.21957600E+00 ????1.00000000 ???
> P ??1
> ?? ???????4.86856000E-01 ????1.00000000 ???
> P ??1
> ?? ???????2.02856000E-01 ????1.00000000 ???
> D ??2
> ?? ???????1.46102300E+00 ????0.01707032 ???
> ?? ???????5.43002000E-01 ????0.10550766 ???
> D ??1
> ?? ???????2.02307000E-01 ????0.00317570 ???
> F ??1
> ?? ???????4.85844740E-01 ????1.00000000 ???
> G ??1
> ?? ???????5.32000000E-01 ????1.00000000 ???

>F ??9.0 ??0.00000 ??0.00000 ??1.98769
> S ??3
> ?? ???????1.15670231E+03 ????0.71431186 ???
> ?? ???????3.45040302E+02 ????1.84554195 ???
> ?? ???????1.11332058E+02 ????4.71095270 ???
> S ??1
> ?? ???????3.85902340E+01 ????4.21329515 ???
> S ??1
> ?? ???????1.42414660E+01 ????0.11772222 ???
> S ??1
> ?? ???????5.53413200E+00 ????0.15894702 ???
> S ??1
> ?? ???????2.23509600E+00 ????1.49637682 ???
> S ??1
> ?? ???????9.24514000E-01 ????0.68056750 ???
> S ??1
> ?? ???????3.85516000E-01 ????0.11297897 ???
> P ??1
> ?? ???????3.12693419E+00 ????1.00000000 ???
> P ??1
> ?? ???????1.26740199E+00 ????1.00000000 ???
> P ??1
> ?? ???????5.28084163E-01 ????1.00000000 ???
> D ??1
> ?? ???????2.31855700E+01 ????0.04106421 ???
> D ??1
> ?? ???????6.68248300E+00 ????0.13327013 ???
> D ??1
> ?? ???????2.16767700E+00 ????0.12350252 ???
> D ??1
> ?? ???????7.33849000E-01 ????0.03195039 ???
> F ??1
> ?? ???????2.55707762E+00 ????1.00000000 ???

> $END
>


Message 4df668afbHW-6706-1288+00.htm, number 365, was edited on Sun May 11 '08 at 21:28:18
and replaces message 4df668afbHW-6706-668+00.htm

Re^2: pcgamess -t switch and parallel runs

Alex Granovsky
gran@classic.chem.msu.su


Hi,

some minor comments to all involved into discussion:

1. -t switch requires one only one argument, the base of the
temporary file dirs. E.g., if it is: -t /tmp/pcg, each PC GAMESS
instance will create directory /tmp/pcg.#nodenumber and will use
it for scratch files. Thus you'll get directories /tmp/pcg.0, tmp/pcg.1,
etc... This resolves any possible conflicts with scratch files.

2. If you are using the alternative syntax with explicit directories
list, they should exist prior to PC GAMESS execution and they should
be chosen so that to avoid any scratch file conflicts - i.e., they
should all be different on the same node

3. Temporary directories should always reside on the local filesystems.

4. MPICH PC GAMESS executable which is dynamically linked is linked
statically against mpich libraries and dynamically against all other,
system libraries. This allow PC GAMESS to run in parallel or in
sequential mode even without installed MPICH at all. Other PC
GAMESS/dynamically linked versions are linked dynamically with MPI
libraries they use.

Regards,
Alex Granovsky







[ This message was edited on Sun May 11 '08 by the author ]


Message 5c715fc0PMq-6707-80-00.htm, number 369, was posted on Mon May 12 '08 at 01:20:24
Needing for sample

Vitaliy Gorbenko
vgorbenko@ukr.net


Have you got a sample of 2-D mapping of PES? Please give it's URL.

Thank you in advance

Vitaliy Gorbenko


Message d4c0ee0b2GM-6707-933+00.htm, number 370, was posted on Mon May 12 '08 at 15:33:04
in reply to 5c715fc0PMq-6707-80-00.htm

Re: Needing for sample

Igor Polyakov
polyakoviv@gmail.com


Hello,
Please see http://classic.chem.msu.su/gran/gamess/scan.html on this website for description of keywords and sample input file.
Good luck!

Message 5b7cb641PMq-6707-1358-00.htm, number 371, was posted on Mon May 12 '08 at 22:37:55
Parameters of $SURF group

Vitaliy Gorbenko
vgorbenko@ukr.net


How many atoms may be included into IGRP-parameter of $SURF group?

Message 5b7cb641PMq-6707-1369+00.htm, number 372, was posted on Mon May 12 '08 at 22:48:57
in reply to d4c0ee0b2GM-6707-933+00.htm

Re^2: Needing for sample

Vitaliy Gorbenko
vgorbenko@ukr.net


On Mon May 12 '08, Igor Polyakov wrote
--------------------------------------
>Hello,
>Please see http://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.html on this website for description of keywords and sample input file.
>Good luck!

But it's URL to 1-Dimensional example! Is there an example with using IVEC1 and IVEC2 simultaneously and with COORD=CART in $CONTROL group?
Thank you!


Message 8afd7dbbaKC-6708-86+00.htm, number 373, was posted on Tue May 13 '08 at 01:26:57
in reply to 4df668afbHW-6706-673+00.htm

Re^2: TDDFT/TDA_restarts

Richard
bonarlaw@liv.ac.uk


Thanks Alex, this might be a useful feature for large molecules, since if I understand it right, these are all-or-nothing calculations i.e. one can't just ask for the high energy states alone.

A related problem: have you any suggestions for when the Davidson iterations show no sign of converging even when using more accurate DFT ? (I used the settings suggested in the May 14 '07 post)

Thanks

Richard



On Sun May 11 '08, Alex Granovsky wrote
---------------------------------------
>Dear Richard,

>while it is impossible at present (just because this is not
>programmed), there should not be any serious problems with
>implementation of this feature. At present, old CIS/TD vectors
>are only used as a starting guess for the subsequent calculations
>at the distorted geometries (e.g., geometry optimization).

>Best regards,
>Alex
>
>
>
>On Fri May 9 '08, Richard wrote
>-------------------------------
>>I don't pretend to understand how these calculations work, but....having run a TDDFT/TDA job with a particular value of NSTATE, should it be possible to rerun the job with a larger value for NSTATE, making use of the already calculated data so the first set of states don't have to be redone i.e. just carry on? It seems the program expects the number of states in the TDVEC group to match NSTATE.

>>thanks

>>Richard


Message d4c0f196bHW-6708-985+00.htm, number 374, was posted on Tue May 13 '08 at 16:25:04
in reply to 8afd7dbbaKC-6708-86+00.htm

Re^3: TDDFT/TDA_restarts

Alex Granovsky
gran@classic.chem.msu.su


Hi Richard,

>A related problem: have you any suggestions for when the Davidson iterations show no sign of converging even when using more accurate DFT ? (I used the settings suggested in the May 14 '07 post)

Could you please send me your sample input file so I'll be able to examine the behavior of PC GAMESS?

Best regards,
Alex Granovsky


Message d4c0f196bHW-6708-987+00.htm, number 375, was posted on Tue May 13 '08 at 16:27:18
in reply to 5b7cb641PMq-6707-1358-00.htm

Re: Parameters of $SURF group

Alex Granovsky
gran@classic.chem.msu.su


Hi,

up to the total number of atoms in system.

Regards,
Alex Granovsky


On Mon May 12 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>How many atoms may be included into IGRP-parameter of $SURF group?


Message 9761ecf13uI-6708-1302-00.htm, number 376, was posted on Tue May 13 '08 at 21:42:46
What compiler for MPICH-Linux

andrea
agalparone@unict.it


Dear PC GAMESS users,
I would be pleased to known what is the compiler version used to
build the parallel Linux binaries linked with MPICH.

Thanks



Message 5c714d31PMq-6709-15+00.htm, number 377, was posted on Wed May 14 '08 at 00:15:37
in reply to 9761ecf13uI-6708-1302-00.htm

Re: What compiler for MPICH-Linux

Vitaliy Gorbenko
vgorbenko@ukr.net


On Tue May 13 '08, andrea wrote
-------------------------------
>Dear PC GAMESS users,
>I would be pleased to known what is the compiler version used to
>build the parallel Linux binaries linked with MPICH.

>Thanks
>
>
>
For both MPICH-1.2.7 and MPICH2-1.0 I've used GCC version 4 from usual set of packages in Fedora Core 8.

Vitaliy Gorbenko


Message d4c0fb6d2GM-6709-157+00.htm, number 378, was posted on Wed May 14 '08 at 02:37:32
in reply to 9761ecf13uI-6708-1302-00.htm

Re: What compiler for MPICH-Linux

Igor Polyakov
polyakoviv@gmail.com



>Dear PC GAMESS users,
>I would be pleased to known what is the compiler version used to
>build the parallel Linux binaries linked with MPICH.

>Thanks

Hello,
If You mean what compiler is used to compile PC GAMESS linked with MPICH/linux the answer is WATCOM for Windows=). This is actually written on the homepage.

Good luck!


Message 4df668afbHW-6714-1203+00.htm, number 379, was posted on Mon May 19 '08 at 20:03:37
in reply to 5b7cb641PMq-6707-1369+00.htm

Re^3: Needing for sample

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I've attached to this posting the sample you asked for.
Note now we have PC GAMESS binaries (available upon request)
which allows one to scan bonds, angles, and torsions while
internally working in DLCs. From now, this is the preferred
way of scan should be performed.

Regards,
Alex Granovsky





On Mon May 12 '08, Vitaliy Gorbenko wrote
-----------------------------------------
>On Mon May 12 '08, Igor Polyakov wrote
>--------------------------------------
>>Hello,
>>Please see http://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.htmlhttp://classic.chem.msu.su/gran/gamess/scan.html on this website for description of keywords and sample input file.
>>Good luck!

>But it's URL to 1-Dimensional example! Is there an example with using IVEC1 and IVEC2 simultaneously and with COORD=CART in $CONTROL group?
>Thank you!

This message contains the 3 kb attachment
[ relaxed_scan_2d_sample_input.txt ] 2D relaxed scan sample


Message 9761ecf63uI-6721-759-00.htm, number 380, was posted on Mon May 26 '08 at 12:39:01
Format of the machinefile

andrea
agalparone@unict.it


Dear PC Gamess users, I would be pleased to known the correct format of the machinefile to run the PCGAMESS MPICH-binary under Linux machines.

Thanks in advance
Andrea


Message d4c0ee0b2GM-6721-936+00.htm, number 381, was posted on Mon May 26 '08 at 15:35:51
in reply to 9761ecf63uI-6721-759-00.htm

Re: Format of the machinefile

Igor Polyakov
polyakoviv@gmail.com


On Mon May 26 '08, andrea wrote
-------------------------------
>Dear PC Gamess users, I would be pleased to known the correct format of the machinefile to run the PCGAMESS MPICH-binary under Linux machines.

>Thanks in advance
> Andrea

Hello!

I use the following for my needs
192.168.10.7:2
192.168.10.8:1
192.168.10.9:1

Which means in general case :. This should work.

Best wishes, Igor.


Message d4c0ee0b2GM-6721-1063+00.htm, number 381, was edited on Mon May 26 '08 at 17:43:24
and replaces message d4c0ee0b2GM-6721-936+00.htm

Re: Format of the machinefile

Igor Polyakov
polyakoviv@gmail.com


On Mon May 26 '08, andrea wrote
-------------------------------
>Dear PC Gamess users, I would be pleased to known the correct format of the machinefile to run the PCGAMESS MPICH-binary under Linux machines.

>Thanks in advance
> Andrea

Hello!

I use the following for my needs
192.168.10.7:2
192.168.10.8:1
192.168.10.9:1

Which means in general case (machine_ip_adress):(number_of_cpu_cores_on this_machine)
This should work.

Best wishes, Igor.

[ This message was edited on Mon May 26 '08 by the author ]


Message 9761ecf63uI-6722-827-00.htm, number 382, was posted on Tue May 27 '08 at 13:46:59
tested MPICH-binary use cases

andrea
agalparone@unict.it


Dear PC Gamess users,
I would be pleased to receive some tested parallel use cases (a part those given at:
http://classic.chem.msu.su/gran/gamess/pcgamess_p4_performance.html and http://classic.chem.msu.su/gran/gamess/p2p.html) for PCGAMESS MPICH-binary running under Linux machines.

Thanks.
Andrea


Message 3edb6f97Ttr-6722-1099-00.htm, number 383, was posted on Tue May 27 '08 at 18:19:45
One level up on TDDFT

shimshon
shimshonkallush@gmail.com


I am using TDDFT and TDHF to calculate two photon properties.
The size of my molecules is around 30-60 heavy atoms, and I commonly use DZV and a basis set.

I calculate transition dipole moments and excitation energies for the first few excited states.

A suitable scaling & benchmarking give satisfactory agreement with experiments.

For some cases, higher theory is needed - probably due to charge transfer on the excited state that demands a use of more then a single reference computation.

I would be grateful for any suggestion regarding any improved method that could be used for this size of molecules.

(maybe some kind of VB will do? could it be combined with TD methods?)

Thank you in advanced,
Shimshon Kallush


Message d4c0fb6d2GM-6723-130+00.htm, number 384, was posted on Wed May 28 '08 at 02:10:00
in reply to 3edb6f97Ttr-6722-1099-00.htm

Re: One level up on TDDFT

Igor Polyakov
polyakoviv@gmail.com


Hello,

I don't know what kinds of molecules u use, but if the active space for excitation is not huge u can try MRMP2 or MCQDPT2 on top of sa-CASSCF.

Good luck!


Message d4c0fb6d2GM-6723-166+00.htm, number 384, was edited on Wed May 28 '08 at 02:46:18
and replaces message d4c0fb6d2GM-6723-130+00.htm

Re: One level up on TDDFT

Igor Polyakov
polyakoviv@gmail.com


Hello,

I don't know what kinds of molecules u are calculating, but if the active space for excitation is not huge u can try MRMP2 or MCQDPT2 on top of sa-CASSCF.

Good luck!

[ This message was edited on Wed May 28 '08 by the author ]


Message 9bcf055e3sn-6724-780-00.htm, number 385, was posted on Thu May 29 '08 at 13:02:15
conversion to natural internal coordinates - intc program help

Kyriakos Kachrimanis
kgk@pharm.auth.gr


Could someone please provide help on the use of the intc program that converts to natural internal coordinates? In particular, it is not clear what keywords can be used in the input file. Another point is how to input the geometry to pcgamess. It would be most helpful if someone could provide an example intc input of a complex organic molecule and the resulting pcgamess input.
Thank you for your time.
Regards

Kyriakos.


Message 93868d6dSol-6726-22-00.htm, number 386, was posted on Sat May 31 '08 at 00:21:56
QM/MM

Jozsef Csontos
jcsontos.lists@gmail.com


Hi,

I've been wondering where I could find documentations/tutorials/examples about the QM/MM capabilities of gamess/pc-gamess.

Is tinker the only option for the MM part?
Are there any differencies between gamess and pc-gamess regarding this issue?

Thanks,
Jozsef


Message 230a4ee2jxY-6727-876-00.htm, number 387, was posted on Sun Jun 1 '08 at 14:37:13
SCF error when direct SCF turned off

Cun Liu
liucun@gmail.com


Dear PCGAMESS users,

I got strange SCF errors when running a hessian job with dirscf turned off. The first SCF step gave right total energy, but it jumped to unreasonably high starting from second SCF step and eventually turned into ****************. No matter diis, shift or soscf option was turned on or off, the problem remained.
The job runs OK when direct SCF turned on. But I need to do morokuma analysis for a cation benzene complex, which doesn't support direct SCF. Any suggestions to fix this problem?

I'm running smp-parallel pcgamess 7.15 on an xeon machine with windows xp. The input and output file were listed below. Thanks!

Regards,
Cun

Input:
$contrl scftyp=rhf maxit=200
runtyp=hessian icharg=0 mult=1
$end
$system mwords=40 timlim=999999 $end
$guess guess=huckel $end
$p2p p2p=.t. dlb=.t. $end
$scf shift=.f. fdiff=.f. diis=.t. dirscf=.f. soscf=.f.
$end
$force method=numeric $end
$data
......
$END

output:
......
-------------------
RHF SCF CALCULATION
-------------------

NUCLEAR ENERGY = 825.3495537595
MAXIT = 200 NPUNCH= 2
EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
DENSITY CONV= 1.00E-06
MEMORY REQUIRED FOR RHF STEP= 663316 WORDS.

ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR VIR. SHIFT DAMPING
1 0 0 -607.180934139 -607.180934139 565.756909694 0.544929167 0.000000000 1.000000000
* * * INITIATING DIIS PROCEDURE * * *
2 1 0 -608.662267534 -1.481333395************** 0.195078757 0.091835782 1.000000000
3 2 0-35619.915257161-35011.252989627 261.095340597************** 0.114665763 1.000000000
4 3 0-35695.793716461 -75.878459300 163.214527610************** 0.023357107 1.000000000
5 4 0-35648.684394061 47.1093224001546.024113859************** 0.023349824 1.000000000
6 5 0-36087.472437124 -438.7880430631313.608801318************** 0.023348316 1.000000000
7 6 0-35758.822949388 328.649487737 223.728930849************** 0.023343642 1.000000000
8 7 0-35799.360253310 -40.537303922 239.985503726************** 0.023343424 1.000000000
9 8 0-35743.969425290 55.390828020 331.725198057************** 0.023343168 1.000000000
10 9 0-35647.869928968 96.099496322 32.468897393************** 0.023342942 1.000000000
11 10 0-35662.402444239 -14.532515271 54.371504156************** 0.023342870 1.000000000
12 11 0-35654.221949521 8.180494717 26.835946382************** 0.023342857 1.000000000
13 12 0-35650.835663122 3.386286399 108.332739652************** 0.023342829 1.000000000
14 13 0-35637.395310970 13.440352152 12.308026012************** 0.023342763 1.000000000
15 14 0-35633.880353348 3.514957622 78.221090906************** 0.023342785 1.000000000
16 15 0-35657.137207674 -23.256854325 33.008887644************** 0.023342737 1.000000000
17 16 0-35664.355439156 -7.218231482 26.120684920************** 0.023342751 1.000000000
18 17 0-35663.915356538 0.440082618 9.275133689************** 0.023342742 1.000000000
19 18 0-35660.726648397 3.188708141 14.397736578************** 0.023342773 1.000000000
20 19 0-35660.644226777 0.082421620 3.563331977************** 0.023342781 1.000000000
21 20 0-35659.620401595 1.023825182 77.029838024************** 0.023342782 1.000000000
22 21 0-35672.111562145 -12.491160550 86.362862047************** 0.023342870 1.000000000
23 22 0-35693.466094769 -21.354532624 7.074762508************** 0.023342888 1.000000000
24 23 0-35693.540419562 -0.074324793 15.198945399************** 0.023342891 1.000000000
25 24 0-35692.243969789 1.296449774 20.353342223************** 0.023342914 1.000000000
26 25 0-35686.855663560 5.388306229 65.150622780************** 0.023342939 1.000000000
27 26 0-35669.038809192 17.816854368 31.285971604************** 0.023342955 1.000000000
28 27 0-35665.167266849 3.871542343 16.088247941************** 0.023342992 1.000000000
29 28 0-35667.775655585 -2.608388736 43.571350163************** 0.023343005 1.000000000
30 29 0-35665.501588122 2.274067464 6.352313357************** 0.023343003 1.000000000
31 30 0-35664.718386059 0.783202063**************************** 0.023343010 1.000000000
32 31 0****************-78312.032736856**************7505.114390197 0.096619211 1.000000000
33 32 0**************** -2378.524897704**************6854.723511083 0.101984779 1.000000000
34 33 0**************** -753.028930972**************2462.590595536 0.173550349 1.000000000
35 34 0**************** -117.292434893**************1963.253845952 0.041213026 1.000000000
36 35 0**************** 109.340963215**************2447.472794121 0.000000000 1.000000000
37 36 0**************** 1197.197395862**************1738.203768326 0.000000000 1.000000000
38 37 0**************** 1581.266651876**************2612.076039910 0.000000000 1.000000000
39 38 0**************** -1718.434505887**************1770.414969550 0.052802787 1.000000000
40 39 0**************** -453.983271217**************1552.034047023 0.000000000 1.000000000
41 40 0**************** 287.083432777**************1470.781528294 0.000000000 1.000000000
42 41 0**************** 92.408507300**************1598.452852809 0.000000000 1.000000000
43 42 0**************** 200.342361688 542.8521282091685.657035248 0.019452097 1.000000000
44 43 0**************** 36.558317511 869.0538751941652.407458421 0.035493935 1.000000000
45 44 0**************** -27.6105394693739.5407126491667.708880685 0.100091815 1.000000000
46 45 0**************** 33.2345793603177.3738419501677.689592535 0.138648858 1.000000000
47 46 0**************** 4.5263582442022.2165941551686.707253817 0.033891720 1.000000000
48 47 0**************** 62.612522506 591.1142489211679.006072582 0.000000000 1.000000000
49 48 0**************** -2.0122509793314.6338126281668.380713411 0.000000000 1.000000000
......


Message 4df668afbHW-6727-1079+00.htm, number 388, was posted on Sun Jun 1 '08 at 17:59:10
in reply to 230a4ee2jxY-6727-876-00.htm

Re: SCF error when direct SCF turned off

Alex Granovsky
gran@classic.chem.msu.su


Hi,

are you using diffuse basis functions and/or getting warning message on the partial linear dependence of your basis set? You should disable Pople integral package and switch to inttyp=hondo if so, as the precision of Pople integral package is not sufficient for diffuse functions with partial linear dependence. Try:

 $contrl inttyp=hondo $end

Regards,
Alex Granovsky


On Sun Jun 1 '08, Cun Liu wrote
-------------------------------
>Dear PCGAMESS users,

>I got strange SCF errors when running a hessian job with dirscf turned off. The first SCF step gave right total energy, but it jumped to unreasonably high starting from second SCF step and eventually turned into ****************. No matter diis, shift or soscf option was turned on or off, the problem remained.
>The job runs OK when direct SCF turned on. But I need to do morokuma analysis for a cation benzene complex, which doesn't support direct SCF. Any suggestions to fix this problem?

>I'm running smp-parallel pcgamess 7.15 on an xeon machine with windows xp. The input and output file were listed below. Thanks!


Message 4df668afbHW-6727-1097+00.htm, number 389, was posted on Sun Jun 1 '08 at 18:16:41
in reply to 3edb6f97Ttr-6722-1099-00.htm

Re: One level up on TDDFT

Alex Granovsky
gran@classic.chem.msu.su


Hi,

at present, GVB code in PC GAMESS cannot be combined with DFT and TD methods.

If the complete active space for MCSCF calculations can be reasonably truncated still capturing all the important physics, you can try CASSCF. The direct SOSCF CASSCF code in PC GAMESS is very efficient (and was very recently considerably improved again) and allows to handle systems with up to, say, 2000 basis functions and reasonable active spaces routinely, especially if running in parallel. The semidirect fitting-based MCQDPT2 code in PC GAMESS is also very efficient and allows to handle problems of virtually any size.

However, if the active space of the reasonable size cannot be selected for your particular systems, you have no ways how to improve your results beyond TDDFT/TDHF.

Best regards,
Alex Granovsky



On Tue May 27 '08, shimshon wrote
---------------------------------
>I am using TDDFT and TDHF to calculate two photon properties.
>The size of my molecules is around 30-60 heavy atoms, and I commonly use DZV and a basis set.

>I calculate transition dipole moments and excitation energies for the first few excited states.

>A suitable scaling & benchmarking give satisfactory agreement with experiments.

>For some cases, higher theory is needed - probably due to charge transfer on the excited state that demands a use of more then a single reference computation.

>I would be grateful for any suggestion regarding any improved method that could be used for this size of molecules.

>(maybe some kind of VB will do? could it be combined with TD methods?)

>Thank you in advanced,
>Shimshon Kallush


Message 4df668afbHW-6727-1102+00.htm, number 390, was posted on Sun Jun 1 '08 at 18:22:33
in reply to 9761ecf63uI-6722-827-00.htm

Re: tested MPICH-binary use cases

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you can freely use the PC GAMESS' standard testcases you mentioned. If you need any assistance converting serial input files into parallel ones with p2p interface enabled, please let me know.

If you'll develop your own or otherwise get interesting testcases please share them with us.

Regards,
Alex Granovsky

On Tue May 27 '08, andrea wrote
-------------------------------
>Dear PC Gamess users,
>I would be pleased to receive some tested parallel use cases (a part those given at:
>http://classic.chem.msu.su/gran/gamess/pcgamess_p4_performance.htmlhttp://classic.chem.msu.su/gran/gamess/pcgamess_p4_performance.html and http://classic.chem.msu.su/gran/gamess/p2p.htmlhttp://classic.chem.msu.su/gran/gamess/p2p.html) for PCGAMESS MPICH-binary running under Linux machines.

>Thanks.
>Andrea
>


Message 18f757e48Nm-6727-1164+00.htm, number 391, was posted on Sun Jun 1 '08 at 19:25:46
in reply to 93868d6dSol-6726-22-00.htm

Re: QM/MM

Jim Kress
jimkress_35@kressworks.com


On Sat May 31 '08, Jozsef Csontos wrote
---------------------------------------
>Hi,

>I've been wondering where I could find documentations/tutorials/examples about the QM/MM capabilities of gamess/pc-gamess.

>Is tinker the only option for the MM part?
>Are there any differencies between gamess and pc-gamess regarding this issue?

>Thanks,
>Jozsef
>

The documentation for the QM/MM version of PC Gamess is given below. Tinker is not an "option". It is a molecular mechanics system that allows the user to select from a wide variety of force fields to be used in the MM part of the QM/MM calculation.

Also, GAMESS-US uses Tinker as its basic Molecular Mechanics system but the input syntax is completely different. It does not include the UFF force field, nor does it include tip4 H2O as a MM entity.

PC GAMESS QM/MM INPUT DOCUMENTATION

(01Aug07 revision)
NOTE: This is version 6.4 of PC GAMESS.

The presence of the $MOLMEC namelist in the pcGAMESS input data stream triggers the QM/MM functionality.

QM/MM Functionality is supported for the following:

Runtyp=OPTIMIZE, ENERGY, GRADIENT, HESSIAN (numerical)
DFT
SCF=RHF, UHF
MPLEVEL=0, 2
COORD=UNIQUE
PTGRP=C1
Internal or general basis sets
Maximum of 3500 Q+C+M total atoms (see below for definitions)

Note: QM/MM functionality will only work for COORD=UNIQUE and symmetry point group C1

QM/MM functionality is provided using an integrated version of the Tinker Molecular Mechanics Package (v3.7) developed by Professor Jay William Ponder of the Washington University School of Medicine. UFF functionality was independently integrated by Jim Kress.

This integration implements version 3.7 of the Tinker Molecular Mechanics (MM) program. NO OTHER VERSION OF TINKER IS SUPPORTED. Tinker is regularly updated by Professor Ponder and his group. New versions of Tinker are not guaranteed (by Professor Ponder, et al) to provide backwards compatibility with older version input (xyz) or parameter (prm) files. Also, we do not guarantee that the MM capabilities integrated into pcGAMESS will be kept current with the Tinker programs as they evolve.

In addition, only the MM parameter files distributed with pcGAMESS will provide proper results. The use of ANY OTHER MM parameter files will not yield correct answers.

As of 17Feb08, the QM/MM version of PCGAMESS will support a maximum of 3500 atoms (=Q+C+M)

$MOLMEC group

This group of parameters provides control over the QM/MM functionality in pcGAMESS. It is required for QM/MM calculations.

TYPE = Is the name of the desired MM Force Field, e.g. UFF, MM3, etc. The default value is TYPE=UFF. The MM parameter filename is taken as TYPE.prm, e.g. for TYPE=UFF the MM parameter file name should be uff.prm. Similarly, for TYPE=MM3, the filename should be MM3.prm.

Note: when running in parallel, the force field parameter file must be located in the same directory with the pcgamess executable on all nodes being used. Also, case matters. The force field name must be in all upper case.

Special note for type=UFF

Atom types 126 and 127 have been altered to allow the inclusion of TIP3 water molecules as UFF MM molecules in QM calculations. Type 126 is HTIP3 and Type 127 is OTIP3.

RDCAP = A logical variable that specifies if the initial cap atom coordinates will be read from the pcGAMESS input file. The default value is .false. The cap coordinates will be in Angstroms. If any cap atoms coordinates are to be read, input for ALL capping atoms must be provided.

This capability is essential if one wants to restart a previous calculation (e.g. an optimization that did not finish) or wants to use the results of a prior QM/MM calculation in a new calculation (e.g. take results from an optimization and calculate the hessian and normal modes).

Within the pcGAMESS input file, the cap geometry information is contained within the $CAPINP group. The cap coordinate data follows a $CAPINP line, one line per cap atom each following the regular GAMESS input format. I.e. name, charge, x, y, z. The name and charge will be ignored but are included for the user's convenience. The cap coordinate data is followed by a line containing $END

MMPRT = A logical variable that specifies if MM type information for each atom (Q,C and M) will be printed with all coordinate and gradient output. If = .false. then the normal GAMESS output will be generated. The default value is .false.

Note: if you set MMPRT=.true. you will break most of the display programs that parse GAMESS input.

Allowing MMPRT=.false. was done to allow QM/MM calculations to generate GAMESS output that will not break the myriad GAMESS parsers out there, e.g. MOLEKEL, gOpenMol, etc. Note: in either case, cap atom info will be printed in the GAMESS output file so geometries parsed from the GAMESS file will include Cap atoms.

CHGKEY = A logical variable that specifies if the atom charges used for MM force fields will be read from the pcGAMESS input file using the $CHGINP group. The default value is .false.

Within the pcGAMESS input file, the atom charge information is contained within the $CHGINP group. The atom charge data follows a $CHGINP line, one line per atom each containing the atom number and charge following the regular GAMESS input format. The charge data is followed by a line containing $END. No blank lines are allowed.

Note: these charges are relevant for both QM and MM atoms. Charge information for the MM representation of the QM atoms should be set, too.

PRTCHG = A logical variable that (when .true.) results in MM charges on all atoms being printed at the beginning of the calculation at the initial geometry. The default value is .false. In the case of UFF, Qeq charges are printed at each new geometry throughout the calculation.

QPOL = THIS IS NOT IMPLEMENTED IN THE CURRENT VERSION.

A logical variable that (when .true.) results in QC atoms being polarized by MM charges. If .false. no polarization occurs. The default value for QPOL is false. QPOL can only be used with the UFF force field with QEQ charges enabled (TYPE=UFF USEQEQ=.true.) or the AMBER force field.

For TYPE=UFF, the following options apply:

READBO = A logical variable that specifies if additional bond order information will be read from the pcGAMESS input file. It is only used if TYPE=UFF. The default value is.false.

For UFF bond order specification is required for all bonds. The program sets all bond orders = 1 for all bonds specified in the input file. If any bonds are not single bonds, the user MUST specify their bond order using READBO.

For READBO=.true., the bond order information follows a $BNDORD line, one line per each bond order specification. Each bond order is specified as i j bond_order where i and j are atom numbers and bond_order is the order of the bond between atoms i and j. The bond order data is followed by a line containing $END

USEQEQ = A logical variable that specifies if QEQ charges are generated. If false, the usual covalent UFF is used. The default value for USEQEQ is false. QEQ charges can only be used with the UFF force field. Use of QEQ is not recommended since it is limited to systems with no overall charge, is very slow to compute and is unstable.

QM/MM functionality requires $DATA geometry specification data be modified as follows:

Geometry is specified as:

Atm_Name Nuc_Chrge X_Coord Y_Coord Z_Coord MM_Type Atom_Type MM_Atom_Conct

1) Atm_Name, Nuc_Chrge, X_Coord, Y_Coord, and Z_Coord have their usual pcGAMESS definitions.

2) MM_Type takes one of three Character values, Q, M, or C.

a)A Q designation indicates pcGAMESS should treat this atom as a regular QM atom

b)A M designation indicates pcGAMESS should treat this atom as a MM atom

c)A C designation indicates pcGAMESS should treat this atom as a 'capping' atom.

3) Atom_Type is the appropriate MM atom type as specified in the TYPE.prm file used in the pcGAMESS/MM calculation.

4) MM_Atom_Conct is a space delimited, integer list of atoms that are directly connected to Atom_Name being specified

5) All atoms to be capped must be explicitly specified. That is done by marking the interface atoms as having a negative entry in the input connectivity line for the interface atom. For example:

C1 6 0.96893124 1.31494824 0.00993246 C 8 -2 5 9 10

In this example, C1 serves as the interface atom between the QM and MM regions. It's connection to QC atom 2 will be assigned one capping hydrogen atom using the appropriate covalent radii for the initial specification of the coordinates for the capping hydrogen.

6) If using general basis input, each cap atom must have a separate basis set specified. For example, if a cap atom is specified as:

Si16 14.0 1.95469 -1.26061 0.01872 C 11 -1 -3 37 38

Then each capping atom (-1 and -3) must have it's capping atom hydrogen basis specified. The input would look like this:

Si16 14.0 1.95469 -1.26061 0.01872 C 11 -1 -3 37 38
S 2
1 4.501800 0.156285
2 0.681444 0.904691
S 1
1 0.151398 1.000000

S 2
1 4.501800 0.156285
2 0.681444 0.904691

S 1
1 0.151398 1.000000

During the input process, a hydrogen atom will be automatically spawned for each capping atom. The spawned H atom will be given a label composed of the input label for the capping atom followed by a _C

Also as a part of the input process, a MM atom will be spawned for each capping atom. The spawned MM atom will be given a label composed of the input label for the capping atom followed by a _B.

An example: Atom C12 is designated as a capping atom. During the input process, a hydrogen atom labelled C12_C will be added to the QC atom list and a MM atom labelled C12_B will be placed in the total atom list. Thus, the capping atom C12 yields 2 elements in the atom list used in the calculation. One is C12_C (the spawned hydrogen atom) and the other is C12_B, a MM atom.

7) In $DATA, no blank cards are required after M atoms

Examples of QM/MM input can be found at the end of this file and in the directory named QMMM sample files

As of 17Feb08, the QM/MM version of PCGAMESS will support a maximum of 3500 atoms (=Q+C+M). Note, this requires that the user have a minimum of 1.5 GB or RAM just to load the program.

---------------------------------------------------------------------


FORCE FIELD PARAMETER SETS

This section is taken from the Tinker v3.7 User Guide. Force Field parameter files will be found in the directory named MM_prm_files. Note: on Linux systems, the force field parameter file names must be all upper case, e.g. UFF.PRM

The TINKER package is distributed with several force field parameter sets, implementing a selection of widely-used literature force fields as well as the TINKER force field currently under construction in the Ponder lab. We try to exactly reproduce the intent of the original authors of our distributed, third-party force fields. In all cases the parameter sets have been validated against literature reports, results provided by the original developers, or calculations made with the authentic programs. With the few exceptions noted below, TINKER calculations can be treated as authentic results from the genuine force fields. A brief description of each parameter set, including some still in preparation and not distributed with the current version, is provided below with lead references:

AMBER.PRM

AMBER-95 parameters for proteins. The nucleic acid parameters are not implemented yet. Note that with their Cornell force field, the Kollman group has devised separate, fully independent partial charge values for each of the N- and C-terminal residues. At present, the terminal residue charges for TINKERs version maintain the correct formal charge, but redistributed somewhat from the Kollman group values. The file reproduces the authentic parm94 set; torsional parameter changes for parm96 are noted in that section of the file.

W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, Jr., D. M. Ferguson, D. C. Spellmeyer, T.

Fox, J. W. Caldwell and P. A. Kollman, A Second Generation Force Field for the Simulation of Proteins,

Nucleic Acids, and Organic Molecules", J. Am. Chem. Soc., 117, 5179-5197 (1995) [PARM94]

P. Kollman, R. Dixon, W. Cornell, T. Fox, C. Chipot and A. Pohorille, The Development/ Application of a Minimalist Organic/Biochemical Molecular Mechanic Force Field using a Combination of ab Initio Calculations and Experimental Data, in Computer Simulation of Biomolecular Systems, W. F. van Gunsteren, P. K. Weiner, A. J. Wilkinson, eds., Volume 3, 83-96 (1997) [PARM96]

G. Moyna, H. J. Williams, R. J. Nachman and A. I. Scott, Conformation in Solution and Dynamics of a Structurally Constrained Linear Insect Kinin Pentapeptide Analogue, Biopolymers, 49, 403-413 (1999) [AIB charges]

W. S. Ross and C. C. Hardin, Ion-Induced Stabilization of the G-DNA Quadruplex: Free Energy Perturbation Studies, J. Am. Chem. Soc., 116, 4363-4366 (1994) [Alkali Metal Ions]

J. Aqvist, Ion-Water Interaction Potentials Derived from Free Energy Perturbation Simulations, J. Phys. Chem., 94, 8021-8024, 1990 [Alkaline Earth Ions, radii adapted for AMBER combining rule]

CHARMM.PRM

CHARMM22 parameters for proteins. Most of the nucleic acid, lipid and small model compound parameters are not yet implemented.

A. D. MacKerell, Jr., J. Wiorkeiwicz-Kuczera and M. Karplus, An All-Atom Empirical Energy Function for the Simulation of Nucleic Acids, J. Am. Chem. Soc., 117, 11946-11975 (1995)

A. D. MacKerrell, Jr., et al., All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins, J. Phys. Chem. B, 102, 3586-3616 (1998)

S. E. Feller, D. Yin, R. W. Pastor and A. D. MacKerell, Jr., Molecular Dynamics Simulation of Unsaturated Lipids at Low Hydration: Parametrization and Comparison with Diffraction Studies, Biophysical Journal, 73, 2269-2279 (1997) [alkenes]

R. H. Stote and M. Karplus, Zinc Binding in Proteins and Solution - A Simple but Accurate

Nonbonded Representation, Proteins, 23, 12-31 (1995) [zinc ion]

MM2.PRM

Full MM2(91) parameters including pi-systems. The anomeric and electronegativity correction terms are not implemented.

N. L. Allinger, "Conformational Analysis. 130. MM2. A Hydrocarbon Force Field Utilizing V1 and V2 Torsional Terms", J. Am. Chem. Soc., 99, 8127-8134 (1977)

J. T. Sprague, J. C. Tai, Y. Yuh and N. L. Allinger, The MMP2 Calculational Method, J. Comput. Chem., 8, 581-603 (1987)

N. L. Allinger, R. A. Kok and M. R. Imam, Hydrogen Bonding in MM2, J. Comput. Chem., 9, 591-595 (1988)

All parameters distributed with TINKER are from the "MM2 (1991) Parameter Set", as provided by N. L. Allinger, University of Georgia

MM3.PRM

Full MM3(99) parameters including pi-systems. The directional hydrogen bonding term is implemented, but the anomeric, electronegativity, Bohlmann correction terms are not implemented.

N. L. Allinger, Y. H. Yuh and J.-H. Lii, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 1, J. Am. Chem. Soc., 111, 8551-8566 (1989)

J.-H. Lii and N. L. Allinger, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 2. Vibrational Frequencies and Thermodynamics, J. Am. Chem. Soc., 111, 8566-8575 (1989)

J.-H. Lii and N. L. Allinger, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 3. The van der Waals Potentials and Crystal Data for Aliphatic and Aromatic Hydrocarbons, J. Am. Chem. Soc., 111, 8576-8582 (1989)

N. L. Allinger, H. J. Geise, W. Pyckhout, L. A. Paquette and J. C. Gallucci, Structures of Norbornane and Dodecahedrane by Molecular Mechanics Calculations (MM3), X-ray Crystallography, and Electron Diffraction, J. Am. Chem. Soc., 111, 1106-1114 (1989) [torsion-stretch]

N. L. Allinger, F. Li and L. Yan, Molecular Mechanics. The MM3 Force Field for Alkenes, J. Comput. Chem., 11, 848-867 (1990)

N. L. Allinger, F. Li, L. Yan and J. C. Tai, Molecular Mechanics (MM3) Calculations on Conjugated Hydrocarbons, J. Comput. Chem., 11, 868-895 (1990)

J.-H. Lii and N. L. Allinger, Directional Hydrogen Bonding in the MM3 Force Field. I, J. Phys. Org. Chem., 7, 591-609 (1994)

J.-H. Lii and N. L. Allinger, Directional Hydrogen Bonding in the MM3 Force Field. II, J. Comput. Chem., 19, 1001-1016 (1998)

All parameters distributed with TINKER were translated from "MM3 PARAMETERS (1999)" as updated on January 17, 1999 and obtained from the above web sites

MM3PRO.PRM

Protein-only version of the MM3 parameters.

J.-H. Lii and N. L. Allinger, The MM3 Force Field for Amides, Polypeptides and Proteins, J. Comput. Chem., 12, 186-199 (1991)

OPLS.PRM

Complete OPLS-UA with united-atom parameters for proteins and many classes of organic molecules. Explicit hydrogens on polar atoms and aromatic carbons.

W. L. Jorgensen and J. Tirado-Rives, The OPLS Potential Functions for Proteins. Energy Minimizations for Crystals of Cyclic Peptides and Crambin, J. Am. Chem. Soc., 110, 1657-1666 (1988) [Protein & Peptide]

W. L. Jorgensen and D. L. Severance, Aromatic-Aromatic Interactions: Free Energy Profiles for the Benzene Dimer in Water, Chloroform, and Liquid Benzene, J. Am. Chem. Soc., 112,4768-4774 (1990) [Aromatic Hydrogens]

S. J. Weiner, P. A. Kollman, D. A. Case, U. C. Singh, C. Ghio, G. Alagona, S. Profeta, Jr. and P. Weiner, A New Force Field for Molecular Mechanical Simulation of Nucleic Acids and Proteins, J. Am. Chem. Soc., 106, 765-784 (1984) [United-Atom "AMBER/OPLS" Local Geometry]

S. J. Weiner, P. A. Kollman, D. T. Nguyen and D. A. Case, An All Atom Force Field for Simulations of Proteins and Nucleic Acids, J. Comput. Chem., 7, 230-252 (1986) [All-Atom "AMBER/OPLS" Local Geometry]

L. X. Dang and B. M. Pettitt, Simple Intramolecular Model Potentials for Water, J. Phys. Chem., 91, 3349-3354 (1987) [Flexible TIP3P and SPC Water]

W. L. Jorgensen, J. D. Madura and C. J. Swenson, Optimized Intermolecular Potential Functions for Liquid Hydrocarbons, J. Am. Chem. Soc., 106, 6638-6646 (1984) [Hydrocarbons]

W. L. Jorgensen, E. R. Laird, T. B. Nguyen and J. Tirado-Rives, Monte Carlo Simulations of Pure Liquid Substituted Benzenes with OPLS Potential Functions, J. Comput. Chem., 14, 206-215 (1993) [Substituted Benzenes]

E. M. Duffy, P. J. Kowalczyk and W. L. Jorgensen, Do Denaturants Interact with Aromatic Hydrocarbons in Water?, J. Am. Chem. Soc., 115, 9271-9275 (1993) [Benzene, Naphthalene, Urea, Guanidinium, TetraMeAmmonium]

W. L. Jorgensen and C. J. Swenson, Optimized Intermolecular Potential Functions for Amides and Peptides. Structure and Properties of Liquid Amides, J. Am. Chem. Soc., 106, 765-784 (1984) [Amides]

W. L. Jorgensen, J. M. Briggs and M. L. Contreras, Relative Partition Coefficients for Organic Solutes form Fluid Simulations, J. Phys. Chem., 94, 1683-1686 (1990) [Chloroform, Pyridine, Pyrazine, Pyrimidine]

J. M. Briggs, T. B. Nguyen and W. L. Jorgensen, Monte Carlo Simulations of Liquid Acetic Acid and Methyl Acetate with the OPLS Potential Functions, J. Phys. Chem., 95, 3315-3322 (1991) [Acetic Acid, Me Acetate]

H. Liu, F. Muller-Plathe and W. F. van Gunsteren, A Force Field for Liquid Dimethyl Sulfoxide and Physical Properties of Liquid Dimethyl Sulfoxide Calculated Using Molecular Dynamics Simulation, J. Am. Chem. Soc., 117, 4363-4366 (1995) [Dimethyl Sulfoxide]

J. Gao, X. Xia and T. F. George, Importance of Bimolecular Interactions in Developing Empirical Potential Functions for Liquid Ammonia, J. Phys. Chem., 97, 9241-9246 (1993) [Ammonia]

J. Aqvist, Ion-Water Interaction Potentials Derived from Free Energy Perturbation Simulations, J. Phys. Chem., 94, 8021-8024 (1990) [Metal Ions]

W. S. Ross and C. C. Hardin, Ion-Induced Stabilization of the G-DNA Quadruplex: Free Energy Perturbation Studies, J. Am. Chem. Soc., 116, 4363-4366 (1994) [Alkali Metal Ions]

J. Chandrasekhar, D. C. Spellmeyer and W. L. Jorgensen, Energy Component Analysis for Dilute Aqueous Solutions of Li+, Na+, F-, and Cl- Ions, J. Am. Chem. Soc., 106, 903-910 (1984) [Halide Ions]

Most parameters distributed with TINKER are from "OPLS and OPLS-AA Parameters for Organic Molecules, Ions, and Nucleic Acids" as provided by W. L. Jorgensen, Yale University, October 1997

OPLSAA.PRM

OPLS-AA with all-atom parameters for proteins and many general classes of organic molecules.

W. L. Jorgensen, D. S. Maxwell and J. Tirado-Rives, Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids, J. Am. Chem. Soc., 117, 11225-11236 (1996)

W. L. Jorgensen and N. A. McDonald, Development of an All-Atom Force Field for Heterocycles. Properties of Liquid Pyridine and Diazenes, THEOCHEM-J. Mol. Struct., 424, 145-155 (1998)

N. A. McDonald and W. L. Jorgensen, Development of an All-Atom Force Field for Heterocycles. Properties of Liquid Pyrrole, Furan, Diazoles, and Oxazoles, J. Phys. Chem. B, 102, 8049-8059 (1998)

All parameters distributed with TINKER are from "OPLS and OPLS-AA Parameters for Organic Molecules, Ions, and Nucleic Acids" as provided by W. L. Jorgensen, Yale University, October 1997

UFF.PRM

UFF all atom parameter set taken from the original work by Rappe and Goddard:

Rappe, CJ Casewit, KS Colwell, WA Goddard III, UFF - A full periodic table force field for molecular mechanics and molecular dynamics simulation, J Am Chem Soc, 1992, 114, 10024-10035

Special note for type=UFF

Atom types 126 and 127 have been altered to allow the inclusion of TIP3 water molecules as UFF MM molecules in QM calculations. Type 126 is HTIP3 and Type 127 is OTIP3.

---------------------------------------------------------------------


This message contains the 58 kb attachment
[ QMMM input documentation 01Aug07.htm ] QM/MM Documentation


Message d4c0fb6d2GM-6727-1326+00.htm, number 392, was posted on Sun Jun 1 '08 at 22:06:27
in reply to 230a4ee2jxY-6727-876-00.htm

Re: SCF error when direct SCF turned off

Igor Polyakov
polyakoviv@gmail.com


Hello!

Actually you can try to supply better initial guess, because sometimes simple huckel is not good enough. For example you can try to compute dft orbitals and use them with guess moread for initial "good" approximation.
Also (as Alex mentioned in the 2nd message of the thread) in the case of partial linear dependence you can use more precise integral evaluation or remove contominants(usually diffuse functions) or use spherical basis functions with d5=.t. set in contrl group. I suppose if you provide the exact input file it will be easier to find the right solution.

Best luck, Igor.


Message 93868d6dSol-6728-1175+00.htm, number 393, was posted on Mon Jun 2 '08 at 19:34:58
in reply to 18f757e48Nm-6727-1164+00.htm

Re^2: QM/MM

Jozsef Csontos
jcsontos.lists@gmail.com


On Sun Jun 1 '08, Jim Kress wrote
---------------------------------

>The documentation for the QM/MM version of PC Gamess is given below.

Thank you very much!

>Tinker is not an "option". It is a molecular mechanics system that allows the user to >select from a wide variety of force fields to be used in the MM part of the QM/MM >calculation.

I'm aware of that. I meant to ask whether there are other MM packages which have been interfaced with pc-gamess. Now, I know the answer:)

Thank you again,

Jozsef




Message 8072e9e8Uag-6743-591-00.htm, number 394, was posted on Tue Jun 17 '08 at 09:51:01
display of TD-DFT transitions/spectrum

Mike
rose@chemistry.ucsc.edu


Hi

after a TD-DFT run, does anyone know a good software for displaying the electronic spectrum (by nm, with adjustable Gaussian broadening) as calculated by the TD-DFT calculation? ideally, it would read the *.out file directly. thanks in advance for any tips.


Message d4c0fb6d2GM-6743-1244+00.htm, number 395, was posted on Tue Jun 17 '08 at 20:43:35
in reply to 8072e9e8Uag-6743-591-00.htm

Re: display of TD-DFT transitions/spectrum

Igor Polyakov
polyakoviv@gmail.com


On Tue Jun 17 '08, Mike wrote
-----------------------------
>Hi

>after a TD-DFT run, does anyone know a good software for displaying the electronic spectrum (by nm, with adjustable Gaussian broadening) as calculated by the TD-DFT calculation? ideally, it would read the *.out file directly. thanks in advance for any tips.
>

Hello!
As I remember chemcraft can display the spectrum, but without Gaussian broadening because u do not compute it. Actually in the outfiles u have the data in nm so there is no need for additional calculations and software.
Best luck.


Message 57db7a95Y6C-6751-807-00.htm, number 396, was posted on Wed Jun 25 '08 at 13:27:25
ERROR, NO MORE THAN 1012 TOTAL ATOMS

Jose A. Teruel
teruel@um.es


Dear PC Gamess users,
I am a new user of PCGAMESS, and I am trying to optimize a large system of about 3600 atoms in a single P4 computer with 4Gb RAM running linux OS and I get *** ERROR, NO MORE THAN 1012 TOTAL ATOMS.

does it mean that I need more RAM?
Anyone has an idea of what is going on?

Thanks a lot
Jose A.


Message 543496aeQ3T-6751-831-00.htm, number 397, was posted on Wed Jun 25 '08 at 13:50:49
Core-hole state geometry opt. and energy

Gregor Kladnik
gregor.kladnik@gmail.com


Hello!

I am a new PC GAMESS user, and i am interested in calculation of core-hole states. Is there an option in the input file for calculating such states? Or is there some other way to do it?

Thank you all for your help!

Gregor Kladnik


Message 459e9a956ZW-6751-1004-00.htm, number 398, was posted on Wed Jun 25 '08 at 16:44:55
Transition metals and input files

Antonio De Crisci
antonio.decrisci@utoronto.ca


Hello All,

I perform various calculations using transition metals with organic molecules attached to them. I normally use the LANL's and SDD's for my ECP's. My question mainly stems from generating input files for these calculations, which sometime can have more than 100 atoms. If for example, if I would like to treat all the organic atoms with 6311G (which is built in PC GAMESS) and than treat the metal with LANL2DZ, do I still have to write every single atom in the input file with the basis set below its coordinates even though 6311G is built into PC GAMESS? Or can I just define the ECP metal in the input file with the ECP basis set of my choice (from the ESML basis set exchange) and the rest can just be assigned to the built in 6311G basis set?

Great program!

Thanks for your time,
Anthony


Message 81706dfbVkC-6751-1091-00.htm, number 399, was posted on Wed Jun 25 '08 at 18:10:40
NaN

kgreen
kayla.green@utsouthwestern.edu


How do i fix: ERROR! NOT 5 OR 6 TR MODES NaN ?

Message 8e01b1486ZW-6751-1153-00.htm, number 400, was posted on Wed Jun 25 '08 at 19:13:34
frequency dependent hyperpolarizibility calculation (NLO) only calculates static values

Antonio De Crisci
antonio.decrisci@utoronto.ca


Hello All,

Im trying to calculate the NLO properties of a molecule at three different wavelengths= 0, 800 and 1029 nm. The calculation proceeds quite nicely, however when the calculation terminates normally, the output file only contains the NLO properties at 0 or the static hyperpolarizilblity, and does not output the values at the other two wavelength. Is there something wrong with my input header? (please see below)

----------------------------------------------------

$BASIS GBASIS=N311 NGAUSS=6 NDFUNC=1 POLAR=POPN311 $END

$CONTRL NOSYM=1 SCFTYP=RHF RUNTYP=TDHF DFTTYP=B3LYP $END

$TDHF NFREQ=3 FREQ(3)=0.000,0.056,0.044 GTHG=.TRUE. MAXITU=9999 BTOL=1.0d-10

ATOL=1.0d-10 MAXITA=9999 $END

$SYSTEM TIMLIM=142560 MEMORY=187976128 MKLNP=4 $END

$SCF DIRSCF=.TRUE. $END

--------------------

Thank you,

Anthony


Message d4c0fb6d2GM-6752-734+00.htm, number 401, was posted on Thu Jun 26 '08 at 12:13:58
in reply to 57db7a95Y6C-6751-807-00.htm

Re: ERROR, NO MORE THAN 1012 TOTAL ATOMS

Igor Polyakov
polyakoviv@gmail.com


On Wed Jun 25 '08, Jose A. Teruel wrote
---------------------------------------
>Dear PC Gamess users,
>I am a new user of PCGAMESS, and I am trying to optimize a large system of about 3600 atoms in a single P4 computer with 4Gb RAM running linux OS and I get *** ERROR, NO MORE THAN 1012 TOTAL ATOMS.

>does it mean that I need more RAM?
>Anyone has an idea of what is going on?

>Thanks a lot
>Jose A.

Hello!

The idea is that the parameter mxatm was set 1012 when PC GAMESS binaries were compiled. Of course if the maximum array dimension is 1012 u cannot store more data and hence u cannot run ur calculation of 3600 atoms. Actually for such a large system u cannot use accurate methods so maybe for the sake of the speed of calculation in this particular case u can switch to simple molecular mechanics programme like tinker or namd. But maybe if u can get PC GAMESS distribution with much higher limits on array dimensions u will be able to run rather fast semiempirical calculations and not get out of memory.

Best luck, Igor


Message d4c0fb6d2GM-6752-740+00.htm, number 402, was posted on Thu Jun 26 '08 at 12:20:13
in reply to 459e9a956ZW-6751-1004-00.htm

Re: Transition metals and input files

Igor Polyakov
polyakoviv@gmail.com


On Wed Jun 25 '08, Antonio De Crisci wrote
------------------------------------------
>Hello All,

>I perform various calculations using transition metals with organic molecules attached to them. I normally use the LANL's and SDD's for my ECP's. My question mainly stems from generating input files for these calculations, which sometime can have more than 100 atoms. If for example, if I would like to treat all the organic atoms with 6311G (which is built in PC GAMESS) and than treat the metal with LANL2DZ, do I still have to write every single atom in the input file with the basis set below its coordinates even though 6311G is built into PC GAMESS? Or can I just define the ECP metal in the input file with the ECP basis set of my choice (from the ESML basis set exchange) and the rest can just be assigned to the built in 6311G basis set?

>Great program!

>Thanks for your time,
>Anthony
>

Hello,

I think actually one of the ways to deal with ur problem is to create external basis set file in which u should define elements like H,C,N,O with 6-311G and ur transition metall with LANL2DZ, name this basis set somehow like "custom" or whatever and then just write in the input file $basis gbasis=custom extfil=.t. $end
U can pick up the custom.lib with basis set definition with the -b switch in the command line for starting PC GAMESS.


Message d4c0fb6d2GM-6752-745+00.htm, number 403, was posted on Thu Jun 26 '08 at 12:24:52
in reply to 81706dfbVkC-6751-1091-00.htm

Re: NaN

Igor Polyakov
polyakoviv@gmail.com


On Wed Jun 25 '08, kgreen wrote
-------------------------------
>How do i fix: ERROR! NOT 5 OR 6 TR MODES NaN ?

Hello,
Could u please specify ur problem more accurate? E.g. give some description of what u are trying to do and when does this message appear? From the first glance it appears to be the hessian matrix corruption, this way u should simply restart ur optimization calculation from the point with lowest energy.( usually one of the last ones)
Best luck.


Message 8afd7ca0aKC-6752-862+00.htm, number 404, was posted on Thu Jun 26 '08 at 14:22:21
in reply to 459e9a956ZW-6751-1004-00.htm

Re: Transition metals and input files

Richard
bonarlaw@liv.ac.uk


Another way, if nobody has told you, is to use the handy include (@) form of input, which also requires you make your own external basis/ecp files - see the input preprocessing example in the readme file section.

Actually, on the subject of specifying ecps, a small query for the developers: how difficult can it be to modify the logic of the parser (or whatever it is that interprets GAMESS-US style input files) so that one only has to list ecps for the elements that require them, rather than all the atoms (and in the right order). A minor thing I know, but it would make inputs simpler for large molecules.

Richard


Message 459e92786ZW-6752-1015+00.htm, number 405, was posted on Thu Jun 26 '08 at 16:55:20
in reply to 459e9a956ZW-6751-1004-00.htm

Re: Transition metals and input files

Antonio De Crisci
antonio.decrisci@utoronto.ca


This is exactly what I was looking for.

Thanks for your replies.

Cheers,
Antonio


Message d42ef48ebHW-6752-1043+00.htm, number 406, was posted on Thu Jun 26 '08 at 17:23:46
in reply to 8afd7ca0aKC-6752-862+00.htm

Re^2: Transition metals and input files

Alex Granovsky
gran@classic.chem.msu.su


Communicated by Roman Zubatyuk

Yet another option - one can create external basis set file containing
also ECPs, and use a script for input file generation. There is one in
Downloads page of PC GAMESS site.

> Another way, if nobody has told you, is to use the handy include (@) form of
> input, which also requires you make your own external basis/ecp files - see
> the input preprocessing example in the readme file section.
>
> Actually, on the subject of specifying ecps, a small query for the developers:
> how difficult can it be to modify the logic of the parser (or whatever it is
> that interprets GAMESS-US style input files) so that one only has to list ecps
> for the elements that require them, rather than all the atoms (and in the
> right order). A minor thing I know, but it would make inputs simpler for large
> molecules.
>
> Richard
>
>
>
>


Message d42eec3cbHW-6752-1173+00.htm, number 407, was posted on Thu Jun 26 '08 at 19:33:43
in reply to 8e01b1486ZW-6751-1153-00.htm

Re: frequency dependent hyperpolarizibility calculation (NLO) only calculates static values

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the correct input lines are:

 $TDHF NFREQ=3 FREQ(1)=0.000,0.056,0.044 GTHG=.TRUE. MAXITU=1000
       BTOL=1.0d-10 ATOL=1.0d-10 MAXITA=1000 
 $END

Note you should start input freq() array at freq(1) index.

Do not use too large MAXITU/MAXITA as it would require extra large amount of memory.

Finally, MKLNP=4 will not speed up calculations of that kind, while running PC GAMESS in parallel on your multicore system (with mklnp=1) do will speed things up.

Best regards,
Alex Granovsky





On Wed Jun 25 '08, Antonio De Crisci wrote
------------------------------------------
>Hello All,

>Im trying to calculate the NLO properties of a molecule at three different wavelengths= 0, 800 and 1029 nm. The calculation proceeds quite nicely, however when the calculation terminates normally, the output file only contains the NLO properties at 0 or the static hyperpolarizilblity, and does not output the values at the other two wavelength. Is there something wrong with my input header? (please see below)

>----------------------------------------------------

>$BASIS GBASIS=N311 NGAUSS=6 NDFUNC=1 POLAR=POPN311 $END

>$CONTRL NOSYM=1 SCFTYP=RHF RUNTYP=TDHF DFTTYP=B3LYP $END

>$TDHF NFREQ=3 FREQ(3)=0.000,0.056,0.044 GTHG=.TRUE. MAXITU=9999 BTOL=1.0d-10

>ATOL=1.0d-10 MAXITA=9999 $END

>$SYSTEM TIMLIM=142560 MEMORY=187976128 MKLNP=4 $END

>$SCF DIRSCF=.TRUE. $END

>--------------------

>Thank you,

>Anthony
>


Message d42eec3cbHW-6752-1176+00.htm, number 408, was posted on Thu Jun 26 '08 at 19:36:11
in reply to 81706dfbVkC-6751-1091-00.htm

Re: NaN

Alex Granovsky
gran@classic.chem.msu.su


Hi,

could you please provide exact input file, PC GAMESS' build #,
target OS and CPU, as well as description of the computer system
used to run this job and exact commandline used to start PC GAMESS?

Thanks!
Alex Granovsky


On Wed Jun 25 '08, kgreen wrote
-------------------------------
>How do i fix: ERROR! NOT 5 OR 6 TR MODES NaN ?


Message d42eec3cbHW-6752-1180+00.htm, number 409, was posted on Thu Jun 26 '08 at 19:39:54
in reply to 543496aeQ3T-6751-831-00.htm

Re: Core-hole state geometry opt. and energy

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you can try either to use rstrct keyword with the
appropriate initial guess orbitals/reordering instructions
or, more generally, very specific kind of casscf.

Please let me know if you need sample input files.

Regards,
Alex Granovsky


On Wed Jun 25 '08, Gregor Kladnik wrote
---------------------------------------
>Hello!

>I am a new PC GAMESS user, and i am interested in calculation of core-hole states. Is there an option in the input file for calculating such states? Or is there some other way to do it?

>Thank you all for your help!

>Gregor Kladnik


Message d42eec3cbHW-6752-1182+00.htm, number 410, was posted on Thu Jun 26 '08 at 19:42:07
in reply to 8afd7ca0aKC-6752-862+00.htm

Re^2: Transition metals and input files

Alex Granovsky
gran@classic.chem.msu.su


Richard,

thanks for your suggestion - it should be easy to implement this feature.

Best regards,
Alex

>Actually, on the subject of specifying ecps, a small query for the developers: how difficult can it be to modify the logic of the parser (or whatever it is that interprets GAMESS-US style input files) so that one only has to list ecps for the elements that require them, rather than all the atoms (and in the right order). A minor thing I know, but it would make inputs simpler for large molecules.

>Richard


Message d42eec3cbHW-6752-1189+00.htm, number 411, was posted on Thu Jun 26 '08 at 19:49:05
in reply to d4c0fb6d2GM-6752-734+00.htm

Re^2: ERROR, NO MORE THAN 1012 TOTAL ATOMS

Alex Granovsky
gran@classic.chem.msu.su


Hi,

As already pointed out by Igor, due to some internal restrictions,
PC GAMESS is limited to approximately 1000 atoms. This limit can
be extended to some degree, however, it is anyway bad idea to run
HF/DFT calculations even with much smaller number of atoms on a
single-CPU workstation - you need to run them in parallel to get
answer in reasonable time.

Regards,
Alex Granovsky

>>Dear PC Gamess users,
>>I am a new user of PCGAMESS, and I am trying to optimize a large system of about 3600 atoms in a single P4 computer with 4Gb RAM running linux OS and I get *** ERROR, NO MORE THAN 1012 TOTAL ATOMS.

>>does it mean that I need more RAM?
>>Anyone has an idea of what is going on?

>>Thanks a lot
>>Jose A.


Message c6d538aedpc-6752-1405-00.htm, number 412, was posted on Thu Jun 26 '08 at 23:25:50
can we find electron affinity

siva
adusumillivishnu@gmail.com


Hello All,

I am a new PC GAMESS user, and i am interested in calculation of electron affinity of a molecule. Can we find that using GAMESS ? Pls send me an exaple if u have any problem related to that.

Thank you all for your help!
Siva.


Message d4c0ee0abHW-6752-1409+00.htm, number 413, was posted on Thu Jun 26 '08 at 23:28:53
in reply to 81706dfbVkC-6751-1091-00.htm

Re: NaN

Alex Granovsky
gran@classic.chem.msu.su


Hi,

thanks for sending me your input and output files.

Actually, you are using GAMESS(US), not PC GAMESS
and thus this is the GAMESS(US) bug.

I was able to successfully run your your input file
with PC GAMESS on my quad-core Core 2 system.
It took about 12 minutes to complete.

Regards,
Alex Granovsky


On Wed Jun 25 '08, kgreen wrote
-------------------------------
>How do i fix: ERROR! NOT 5 OR 6 TR MODES NaN ?


Message 8e9608756ZW-6752-1435+00.htm, number 414, was posted on Thu Jun 26 '08 at 23:55:48
in reply to 459e9a956ZW-6751-1004-00.htm

Re: Transition metals and input files

Antonio De Crisci
antonio.decrisci@utoronto.ca



So just a follow up. I made an external file which defines the basis functions for each of the different atoms in my molecule instead of EACH atom. This comes in handy especially if there is an interruption in the calculation.

Here is my input header:

$BASIS EXTFIL=.TRUE. GBASIS=CUSTOMPT $END
$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP MAXIT=9999 $END
$SYSTEM TIMLIM=142560 MEMORY=187976128 $END
$GUESS PRTMO=.TRUE. $END
$SCF DIRSCF=.TRUE. $END
$STATPT OPTTOL=0.0001 NSTEP=9999 STPT=.TRUE. $END

and just below this are the coordinates to my atoms and the ECP section. I ran the calculation (not to completion) and everything proceeds smoothly, no error messages at all and the out file re-writes the basis set for each atom so can double check to see I did everything correctly.

However, what about the ECP section? I have placed the ECP section in the input file just below the $DATA section after the $END as:

$ECP
Pt-ECP GEN 60 5
1 ----- H POTENTIAL -----
0.000000000 2 1.000000000
2 ----- S-H POTENTIAL -----
579.223861000 2 13.428651000
29.669491000 2 6.714326000
2 ----- P-H POTENTIAL -----
280.860774000 2 10.365944000
26.745382000 2 5.182972000
2 ----- D-H POTENTIAL -----
120.396444000 2 7.600479000
15.810921000 2 3.800240000
1 ----- F-H POTENTIAL -----
24.314376000 2 3.309569000
1 ----- G-H POTENTIAL -----
-24.218675000 2 5.277289000
C-ECP NONE
H-ECP NONE
P-ECP NONE
$END

I ask this because in the output file, there is no mention of an ECP being used and I just want to make that it is using ECP on Pt only. The extfil has the Pt basis set, but not the ECP values. Should the ECP section be in the extfil?

As it stands now, this is a working input and I just wanted to make sure its applying ECP to Pt.

Thanks for your time.

Anthony


Message d4c0fb6d2GM-6753-112+00.htm, number 415, was posted on Fri Jun 27 '08 at 01:52:14
in reply to c6d538aedpc-6752-1405-00.htm

Re: can we find electron affinity

Igor Polyakov
polyakoviv@gmail.com


Hello,

Yep, u definitely can do it, but sometimes it can be tricky. The idea is to calculate energy of the regular molecule and then simply add 1 electron by decreasing charge by -1 (icharg in $contrl) and increasing multiplicity by 1 (mult in $contrl). For my particular interests i use mrmp2 on top of casscf to evaluate delta energies between regular molecule and a molecule with electron. I suppose sometimes it is ok to use regular mp2.

Best luck.


Message c2f939e8Q3T-6753-672+00.htm, number 416, was posted on Fri Jun 27 '08 at 11:11:56
in reply to d42eec3cbHW-6752-1180+00.htm

Re^2: Core-hole state geometry opt. and energy

Gregor Kladnik
gregor.kladnik@gmail.com


Hi!

First of all, thanks for the swift answer! Hm, I would really appreciate if you could send some sample input files. Thanks Alex!

Sincerely,
Gregor Kladnik

>Hi,

>you can try either to use rstrct keyword with the
>appropriate initial guess orbitals/reordering instructions
>or, more generally, very specific kind of casscf.

>Please let me know if you need sample input files.

>Regards,
>Alex Granovsky


Message 8afd7ca0aKC-6753-1093-00.htm, number 417, was posted on Fri Jun 27 '08 at 18:13:47
Getting the same answer from US and PC_GAMESS DFT optimisations

Richard
bonarlaw@liv.ac.uk


I want to run a dft geometry optimisation of a smallish organometallic with gamess-us (Apr 11, 2008) so that I get the same answer as using pc-gamess. Setting NRAD=63 in gamess-us, with all other parameters in both programs default (input files identical), I get a difference of ca. 0.01 angs in an interesting bond length, which is a bit worrying.

I realize there are many things to consider here, but does anyone have suggestions for settings in gamess-us which are most likely to be responsible for the discrepancy. Or is it in general possible to do this sort of comparison with different dft codes?

BTW, I'm not interested in timings - I know the answer to that - I just want to end up with optimised structures which are to all intents and purposes the same (I'd be happy with 0.001 angs difference).

Richard




Message d4c0fb6d2GM-6753-1390+00.htm, number 418, was posted on Fri Jun 27 '08 at 23:10:18
in reply to c6d538aedpc-6752-1405-00.htm

Re: can we find electron affinity

Igor Polyakov
polyakoviv@gmail.com


On Thu Jun 26 '08, siva wrote
-----------------------------
>Hello All,

>I am a new PC GAMESS user, and i am interested in calculation of electron affinity of a molecule. Can we find that using GAMESS ? Pls send me an exaple if u have any problem related to that.

>Thank you all for your help!
>Siva.

>

Hello!

Yep, u definitely can do it, but sometimes it can be tricky. The idea is to
calculate energy of the regular molecule and then simply add 1 electron by
decreasing charge by -1 (icharg in $contrl) and increasing multiplicity by 1
(mult in $contrl). For my particular interests i use mrmp2 on top of casscf to
evaluate delta energies between regular molecule and a molecule with electron.
I suppose sometimes it is ok to use regular mp2.

Best luck.


Message d4c0fb6d2GM-6753-1414+00.htm, number 419, was posted on Fri Jun 27 '08 at 23:34:41
in reply to 8afd7ca0aKC-6753-1093-00.htm

Re: Getting the same answer from US and PC_GAMESS DFT optimisations

Igor Polyakov
polyakoviv@gmail.com


Hello,

Maybe theae are some dumb questions from my side, but are u using the same convergence criteria (e.g. opttol) in both pc gamess and gamess(us)? Are u using the same coordinates(cartesian, internal or dlc) in both cases? Are u using exactly the same DFT functional for both tasks?

Best luck.


Message d4c0fb6d2GM-6754-64+00.htm, number 419, was edited on Sat Jun 28 '08 at 01:03:57
and replaces message d4c0fb6d2GM-6753-1414+00.htm

Re: Getting the same answer from US and PC_GAMESS DFT optimisations

Igor Polyakov
polyakoviv@gmail.com


Hello,

Maybe these are some dumb questions from my side, but are u using the same convergence criteria (e.g. opttol) in both pc gamess and gamess(us)? Are u using the same coordinates(cartesian, internal or dlc) in both cases? Are u using exactly the same DFT functional for both tasks?

Best luck.

[ This message was edited on Sat Jun 28 '08 by the author ]


Message 8afd7ca0aKC-6757-1086+00.htm, number 420, was posted on Tue Jul 1 '08 at 18:06:59
in reply to 8afd7ca0aKC-6753-1093-00.htm

Re: Getting the same answer from US and PC_GAMESS DFT optimisations: solved

Richard
bonarlaw@liv.ac.uk


It turns out that for our molecules with fairly soft metal-ligand bonds the default grid settings in GAMESS-US are unreliable. The way to get the same answer as PC-GAMESS (which is the right answer) is to increase the grid size in GAMESS-US using:

$DFT NRAD=128 NLEB=770 $END

Its probably a good idea to tighten OPTTOL to 1e-5 or 1e-6, and check different optimization methods (cartesian etc.) as well.

Richard


Message 4df668afbHW-6757-1305+00.htm, number 421, was posted on Tue Jul 1 '08 at 21:45:31
in reply to c2f939e8Q3T-6753-672+00.htm

Re^3: Core-hole state geometry opt. and energy

Alex Granovsky
gran@classic.chem.msu.su


Hi,

First, I'm sorry for the delay with my answer.

I'm attaching archive with two sample inputs and outputs -
the first one using ROHF, and the second using specific case
of CASSCF. The second is more general and flexible approach.
The model system is core-hole state of water cation.

Please let us know if you have any questions.

Regards,
Alex Granovsky


On Fri Jun 27 '08, Gregor Kladnik wrote
---------------------------------------
>Hi!

>First of all, thanks for the swift answer! Hm, I would really appreciate if you could send some sample input files. Thanks Alex!

>Sincerely,
>Gregor Kladnik

>>Hi,

>>you can try either to use rstrct keyword with the
>>appropriate initial guess orbitals/reordering instructions
>>or, more generally, very specific kind of casscf.

>>Please let me know if you need sample input files.

>>Regards,
>>Alex Granovsky
>

This message contains the 22 kb attachment
[ hole_samples.rar ] Core-hole state of water cation optimization sample


Message 4df668afbHW-6757-1312+00.htm, number 422, was posted on Tue Jul 1 '08 at 21:51:47
in reply to 8e9608756ZW-6752-1435+00.htm

Re^2: Transition metals and input files

Alex Granovsky
gran@classic.chem.msu.su


Hi!

First, you still need to specify ecp=read in the $contrl group.
Second, in your $ECP group you must specify ECPs for all atoms.
Finally, to use ECPs with H potentials, you need to add the following to your input:

 $ecpinp len1=200000 len2=20000 $end

Best regards,
Alex Granovsky





On Thu Jun 26 '08, Antonio De Crisci wrote
------------------------------------------
>>So just a follow up. I made an external file which defines the basis functions for each of the different atoms in my molecule instead of EACH atom. This comes in handy especially if there is an interruption in the calculation.

>Here is my input header:

> $BASIS EXTFIL=.TRUE. GBASIS=CUSTOMPT $END
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP MAXIT=9999 $END
> $SYSTEM TIMLIM=142560 MEMORY=187976128 $END
> $GUESS PRTMO=.TRUE. $END
> $SCF DIRSCF=.TRUE. $END
> $STATPT OPTTOL=0.0001 NSTEP=9999 STPT=.TRUE. $END

>and just below this are the coordinates to my atoms and the ECP section. I ran the calculation (not to completion) and everything proceeds smoothly, no error messages at all and the out file re-writes the basis set for each atom so can double check to see I did everything correctly.

>However, what about the ECP section? I have placed the ECP section in the input file just below the $DATA section after the $END as:

>$ECP
>Pt-ECP GEN 60 5
> 1 ----- H POTENTIAL -----
> 0.000000000 2 1.000000000
> 2 ----- S-H POTENTIAL -----
> 579.223861000 2 13.428651000
> 29.669491000 2 6.714326000
> 2 ----- P-H POTENTIAL -----
> 280.860774000 2 10.365944000
> 26.745382000 2 5.182972000
> 2 ----- D-H POTENTIAL -----
> 120.396444000 2 7.600479000
> 15.810921000 2 3.800240000
> 1 ----- F-H POTENTIAL -----
> 24.314376000 2 3.309569000
> 1 ----- G-H POTENTIAL -----
> -24.218675000 2 5.277289000
>C-ECP NONE
>H-ECP NONE
>P-ECP NONE
>$END

>I ask this because in the output file, there is no mention of an ECP being used and I just want to make that it is using ECP on Pt only. The extfil has the Pt basis set, but not the ECP values. Should the ECP section be in the extfil?

>As it stands now, this is a working input and I just wanted to make sure its applying ECP to Pt.

>Thanks for your time.

>Anthony


Message 4df668afbHW-6757-1314+00.htm, number 423, was posted on Tue Jul 1 '08 at 21:53:56
in reply to d4c0fb6d2GM-6753-1390+00.htm

Re^2: can we find electron affinity

Alex Granovsky
gran@classic.chem.msu.su


Hi,

yes it is often fine with MP2 (and sometimes even with DFT).
Note you must use high-quality basis set with diffuse functions.

Regards,
Alex Granovsky


On Fri Jun 27 '08, Igor Polyakov wrote
--------------------------------------
>On Thu Jun 26 '08, siva wrote
>-----------------------------
>>Hello All,

>>I am a new PC GAMESS user, and i am interested in calculation of electron affinity of a molecule. Can we find that using GAMESS ? Pls send me an exaple if u have any problem related to that.

>>Thank you all for your help!
>>Siva.

>>

>Hello!

>Yep, u definitely can do it, but sometimes it can be tricky. The idea is to
>calculate energy of the regular molecule and then simply add 1 electron by
>decreasing charge by -1 (icharg in $contrl) and increasing multiplicity by 1
>(mult in $contrl). For my particular interests i use mrmp2 on top of casscf to
>evaluate delta energies between regular molecule and a molecule with electron.
>I suppose sometimes it is ok to use regular mp2.

>Best luck.


Message 459e9b866ZW-6758-180+00.htm, number 424, was posted on Wed Jul 2 '08 at 03:00:19
in reply to 4df668afbHW-6757-1312+00.htm

Re^3: Transition metals and input files

Antonio De Crisci
antonio.decrisci@utoronto.ca



Great. This worked!

There is now an ECP section in out file during the calculation.

Thank you,
Anthony

On Tue Jul 1 '08, Alex Granovsky wrote
--------------------------------------
>Hi!

>First, you still need to specify ecp=read in the $contrl group.
>Second, in your $ECP group you must specify ECPs for all atoms.
>Finally, to use ECPs with H potentials, you need to add the following to your input:

>

 $ecpinp len1=200000 len2=20000 $end

>Best regards,
>Alex Granovsky
>
>
>
>
>
>On Thu Jun 26 '08, Antonio De Crisci wrote
>------------------------------------------
>>>So just a follow up. I made an external file which defines the basis functions for each of the different atoms in my molecule instead of EACH atom. This comes in handy especially if there is an interruption in the calculation.

>>Here is my input header:

>> $BASIS EXTFIL=.TRUE. GBASIS=CUSTOMPT $END
>> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=B3LYP MAXIT=9999 $END
>> $SYSTEM TIMLIM=142560 MEMORY=187976128 $END
>> $GUESS PRTMO=.TRUE. $END
>> $SCF DIRSCF=.TRUE. $END
>> $STATPT OPTTOL=0.0001 NSTEP=9999 STPT=.TRUE. $END

>>and just below this are the coordinates to my atoms and the ECP section. I ran the calculation (not to completion) and everything proceeds smoothly, no error messages at all and the out file re-writes the basis set for each atom so can double check to see I did everything correctly.

>>However, what about the ECP section? I have placed the ECP section in the input file just below the $DATA section after the $END as:

>>$ECP
>>Pt-ECP GEN 60 5
>> 1 ----- H POTENTIAL -----
>> 0.000000000 2 1.000000000
>> 2 ----- S-H POTENTIAL -----
>> 579.223861000 2 13.428651000
>> 29.669491000 2 6.714326000
>> 2 ----- P-H POTENTIAL -----
>> 280.860774000 2 10.365944000
>> 26.745382000 2 5.182972000
>> 2 ----- D-H POTENTIAL -----
>> 120.396444000 2 7.600479000
>> 15.810921000 2 3.800240000
>> 1 ----- F-H POTENTIAL -----
>> 24.314376000 2 3.309569000
>> 1 ----- G-H POTENTIAL -----
>> -24.218675000 2 5.277289000
>>C-ECP NONE
>>H-ECP NONE
>>P-ECP NONE
>>$END

>>I ask this because in the output file, there is no mention of an ECP being used and I just want to make that it is using ECP on Pt only. The extfil has the Pt basis set, but not the ECP values. Should the ECP section be in the extfil?

>>As it stands now, this is a working input and I just wanted to make sure its applying ECP to Pt.

>>Thanks for your time.

>>Anthony


Message 57c265d24el-6758-1036-00.htm, number 425, was posted on Wed Jul 2 '08 at 17:16:28
Rigid Body PES Scans

Adam Sykes
adam.sykes@liv.ac.uk


I am trying to perform a calculation in which two water molecules are held in rigid conformations, i.e. HOH angle and OH lengths fixed, but the two water molecules are allowed to move relative to each other.

The calculation should then optimized the orientation and distance between the two bodies.

I can't seem to get the correct constraints setup. I have tried frezing one with IFRZAT(1) and fixing the bonds and angles of the other using IFZMAT(1), but with little success. I want to avoid z-matrices, and would prefer DLC coords if possible.

This calculation may also be expanded to a larger interacting body with water, so avoiding fixing all bonds, angles and dihedrals, instead of fixing atomic positions of the larger body, would be preferable. This would then be used for optimizing partial charges via water interactions, for CHARMM parameterization.

Thank you for your help,

Adam Sykes


Message 9761ecf63uI-6758-1225-00.htm, number 426, was posted on Wed Jul 2 '08 at 20:24:47
Increasing linear dependence threshold when using diffuse functions

andrea
agalparone@unict.it


Dear PC-GAMESS users,
I would be pleased to known if there is a command to increase the linear dependence threshold to achieve SCF convergence, similar to QMTTOL in GAMESS-US version.
Thanks very much.


Message c121ca7baGH-6758-1388+00.htm, number 427, was posted on Wed Jul 2 '08 at 23:07:48
in reply to 9761ecf63uI-6758-1225-00.htm

Re: Increasing linear dependence threshold when using diffuse functions

Roman Zubatyuk
zubatyuk@gmail.com


Dear Andrea,

As far as I know PC GAMESS doesn't have this functionality. It could not drop some functions in order to reduce linear dependence. So, in the case of diffuse basis sets and/or large molecules you need to increase ICUT in $CONTRL (look at http://classic.chem.msu.su/gran/gamess/inttyp.html). For of MP2 calculations you also need to increase TOL1 and TOL2 in $MP2GRD. You need to be careful while doing MP2 with diffuse basis set: if integral thresholds will not be tight enough you have a chance to get wrong result!

On Wed Jul 2 '08, andrea wrote
------------------------------
>Dear PC-GAMESS users,
>I would be pleased to known if there is a command to increase the linear dependence threshold to achieve SCF convergence, similar to QMTTOL in GAMESS-US version.
>Thanks very much.

> �


Message 459e96376ZW-6760-496-00.htm, number 428, was posted on Fri Jul 4 '08 at 08:16:21
fluorescence properties

Antonio De Crisci
antonio.decrisci@utoronto.ca


Hello All,

Im interested in calculating fluorescence properties of molecules. Would it be possible to calculate emission life-time and quantum yield/efficiency of single molecules in PC GAMESS?

Thanks,

Antonio De Crisci
University of Toronto


Message 5c2408102GM-6760-756+00.htm, number 429, was posted on Fri Jul 4 '08 at 12:36:19
in reply to 459e96376ZW-6760-496-00.htm

Re: fluorescence properties

Igor Polyakov
polyakoviv@gmail.com


On Fri Jul 4 '08, Antonio De Crisci wrote
-----------------------------------------
>Hello All,

>Im interested in calculating fluorescence properties of molecules. Would it be possible to calculate emission life-time and quantum yield/efficiency of single molecules in PC GAMESS?

>Thanks,

>Antonio De Crisci
>University of Toronto

Hello,

You can calculate energies of S0-Sn,Sn-S0 transitions, oscillator strength, transitions dipoles. Also u can explore excited state PESs in order to find radiationless chanels to ground state like avoided crossings or minimal energy conical intersection points. All these options are available and are rather fast to calculate in PC GAMESS. The only thing u cannot do - dynamical simulations. I suppose that currently carrying out dynamical simulations of excited states of medium (and larger) size chromophores at a good level of theory is not possible in any computer programme in terms of CPU time.

Best luck.


Message 4df668afbHW-6765-849+00.htm, number 430, was posted on Wed Jul 9 '08 at 14:09:14
in reply to 459e96376ZW-6760-496-00.htm

Re: fluorescence properties

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you can calculate dipole transition moments
and then Einstein coefficients which are
related with spontaneous emission lifetimes, see e.g.,
http://scienceworld.wolfram.com/physics/EinsteinCoefficients.html

However, there are usually multiple different factors
which affect value of quantum yield, so that you need to
have some model of the various possible processes taking
place upon excitation in order to be able to estimate
quantum yield.

Regards,
Alex Granovsky



On Fri Jul 4 '08, Antonio De Crisci wrote
-----------------------------------------
>Hello All,

>Im interested in calculating fluorescence properties of molecules. Would it be possible to calculate emission life-time and quantum yield/efficiency of single molecules in PC GAMESS?

>Thanks,

>Antonio De Crisci
>University of Toronto


Message 4df668afbHW-6765-862+00.htm, number 431, was posted on Wed Jul 9 '08 at 14:22:23
in reply to c121ca7baGH-6758-1388+00.htm

Re^2: Increasing linear dependence threshold when using diffuse functions

Alex Granovsky
gran@classic.chem.msu.su


Dear Andrea,
Dear Roman,

Yes the PC GAMESS does not support QMTTOL option at present.
There are two reasons why it is not implemented:

1. To avoid sudden changes of the number of basis functions
used in calculations like geometry optimization, etc...
The fully automated approach could skip different number
of linear combinations at different points of PES,
resulting in discontinuous PES.

2. To avoid minor errors in energy gradients and hessians
introduced by the use of QMTTOL.

However, PC GAMESS lists the most linearly dependent basis
functions and allows deleting them manually. We are planning
to automate this to some degree in future releases, i.e., to
introduce user-convenient way of how to delete them from
calculations.

Best regards,
Alex Granovsky


On Wed Jul 2 '08, Roman Zubatyuk wrote
--------------------------------------
>Dear Andrea,

> As far as I know PC GAMESS doesn't have this functionality. It could not drop some functions in order to reduce linear dependence. So, in the case of diffuse basis sets and/or large molecules you need to increase ICUT in $CONTRL (look at http://classic.chem.msu.su/gran/gamess/inttyp.html). For of MP2 calculations you also need to increase TOL1 and TOL2 in $MP2GRD. You need to be careful while doing MP2 with diffuse basis set: if integral thresholds will not be tight enough you have a chance to get wrong result!

>On Wed Jul 2 '08, andrea wrote
>------------------------------
>>Dear PC-GAMESS users,
>>I would be pleased to known if there is a command to increase the linear dependence threshold to achieve SCF convergence, similar to QMTTOL in GAMESS-US version.
>>Thanks very much.

>> �


Message 4df668afbHW-6765-866+00.htm, number 432, was posted on Wed Jul 9 '08 at 14:25:37
in reply to 57c265d24el-6758-1036-00.htm

Re: Rigid Body PES Scans

Alex Granovsky
gran@classic.chem.msu.su


Dear Adam,

I'm sorry for delay with my response.

In fact, there are several ways how you can do this with PC GAMESS.

Please let me know if you are still interested in this and I'll attach several samples for you to this posting.


Regards,
Alex Granovsky



On Wed Jul 2 '08, Adam Sykes wrote
----------------------------------
>I am trying to perform a calculation in which two water molecules are held in rigid conformations, i.e. HOH angle and OH lengths fixed, but the two water molecules are allowed to move relative to each other.

>The calculation should then optimized the orientation and distance between the two bodies.

>I can't seem to get the correct constraints setup. I have tried frezing one with IFRZAT(1) and fixing the bonds and angles of the other using IFZMAT(1), but with little success. I want to avoid z-matrices, and would prefer DLC coords if possible.

>This calculation may also be expanded to a larger interacting body with water, so avoiding fixing all bonds, angles and dihedrals, instead of fixing atomic positions of the larger body, would be preferable. This would then be used for optimizing partial charges via water interactions, for CHARMM parameterization.

>Thank you for your help,

>Adam Sykes


Message 4df668afbHW-6765-1029+00.htm, number 433, was posted on Wed Jul 9 '08 at 17:10:01
in reply to 57c265d24el-6758-1036-00.htm

Re: Rigid Body PES Scans

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I'm attaching to this posting archive with three sample input files.

Regards,
Alex Granovsky


On Wed Jul 2 '08, Adam Sykes wrote
----------------------------------
>I am trying to perform a calculation in which two water molecules are held in rigid conformations, i.e. HOH angle and OH lengths fixed, but the two water molecules are allowed to move relative to each other.

>The calculation should then optimized the orientation and distance between the two bodies.

>I can't seem to get the correct constraints setup. I have tried frezing one with IFRZAT(1) and fixing the bonds and angles of the other using IFZMAT(1), but with little success. I want to avoid z-matrices, and would prefer DLC coords if possible.

>This calculation may also be expanded to a larger interacting body with water, so avoiding fixing all bonds, angles and dihedrals, instead of fixing atomic positions of the larger body, would be preferable. This would then be used for optimizing partial charges via water interactions, for CHARMM parameterization.

>Thank you for your help,

>Adam Sykes

This message contains the 2 kb attachment
[ two_waters.rar ] Two rigid body fragments optimization


Message ca6a102439w-6766-287-00.htm, number 434, was posted on Thu Jul 10 '08 at 04:46:51
Questions about relative velocity calculation

Yubing.Si
yubing.si@gmail.com


Dear all,

I am a new user of pc-gamess, some questions puzzled me.

How can i use pc-gamess to calculate the relative velocity of the reactants/transition states at different temperatures? I saw the semi-empirical DRC calculation give the KINETIC ENERGY and VELOCITY in exam27, but how can i definition the Keyword of VEL in this calculation?

Thank you in advance.

Yubing.Si


Message c9524c90a6T-6767-1074-00.htm, number 435, was posted on Fri Jul 11 '08 at 17:54:29
PCM calculate

Rilvia
rilvia@feq.unicamp.br


I'm using this input to calculate volume and area:

$CONTRL UNITS=ANGS COORD=CART RUNTYP=ENERGY SCFTYP=RHF
ICHARG= 0 MULT= 1 $END
$GUESS GUESS= HUCKEL $END
$SYSTEM memory=12000000 $END
$BASIS GBASIS=N31 NGAUSS=6 POLAR=POPLE $END
$PCM SOLVNT=WATER ICAV=1 IDISP=1 $END

$DATA

But in the output show me many volume and area. I would like to know what the diferences among their.

Thaks
Rilvia


Message 18bfebbebrC-6768-106-00.htm, number 436, was posted on Sat Jul 12 '08 at 01:46:34
molecule size

Peter Grotticelli
pgrott22@optonline.net


What's up, my friends? I hail from Stony Brook, NY; and I'm an undergraduate trying simply to geometrically optimize a professor's library.

I ran through GAMESS a twenty-seven-atom molecule and a seventy-atom molecule, each with the same parameters, including C1 (no symmetry). For the smaller molecule I get coordinates for all twenty-seven atoms; but for the larger molecule I get coordinates for only twenty-five random atoms. I have varied the parameters for the larger molecule, and have thus varied the time taken tenfold, but still I get exactly twenty-five atoms, and it is oddly considered to have terminated normally. So it cant be simply a system overload. What might be different about the seventy-atom molecule?

Thanks for any guidance.
- Peter Grotticelli


Message 589de40fbQn-6768-749+00.htm, number 437, was posted on Sat Jul 12 '08 at 12:29:17
in reply to 18bfebbebrC-6768-106-00.htm

Re: molecule size

Pedro Silva
pedros@ufp.pt


Please send us the input file for the larger molecule, so that we can analyze it..

Pedro


Message 18bfebbebrC-6768-1066+00.htm, number 439, was posted on Sat Jul 12 '08 at 17:46:24
in reply to 18bfebbebrC-6768-1065+00.htm

Re^2: molecule size

Peter Grotticelli
pgrott22@optonline.net


On Sat Jul 12 '08, Peter Grotticelli wrote
------------------------------------------
>On Sat Jul 12 '08, Pedro Silva wrote
>------------------------------------
>>Please send us the input file for the larger molecule, so that we can analyze it..

>>Pedro

Here it is. Perhaps the ridiculously high OPTTOL is causing the problem; however, I suspect that it's more general than that.

Thanks.
-Peter

This message contains the 4 kb attachment
[ 70atominput.txt ]


Message 18bfebbebrC-6768-1086+00.htm, number 439, was edited on Sat Jul 12 '08 at 18:05:50
and replaces message 18bfebbebrC-6768-1066+00.htm

Re^2: molecule size

Peter Grotticelli
pgrott22@optonline.net


On Sat Jul 12 '08, Peter Grotticelli wrote
------------------------------------------
>On Sat Jul 12 '08, Pedro Silva wrote
>------------------------------------
>>Please send us the input file for the larger molecule, so that we can analyze it..

>>Pedro

Here it is. Perhaps the ridiculously high OPTTOL is causing the problem; however, I suspect that it's more general than that, especially because I used the same OPTTOL successfully on the 27-atom molecule.

Thanks.
-Peter

This message contains the 4 kb attachment
[ 70atominput.txt ]

[ This message was edited on Sat Jul 12 '08 by the author ]


Message d42eec2dbHW-6768-1168+00.htm, number 440, was posted on Sat Jul 12 '08 at 19:28:27
in reply to 18bfebbebrC-6768-1086+00.htm

Re^3: molecule size

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there are several mistakes in your input, I've attached the corrected one to this posting.

Regards,
Alex


On Sat Jul 12 '08, Peter Grotticelli wrote
------------------------------------------
>On Sat Jul 12 '08, Peter Grotticelli wrote
>------------------------------------------
>>On Sat Jul 12 '08, Pedro Silva wrote
>>------------------------------------
>>>Please send us the input file for the larger molecule, so that we can analyze it..

>>>Pedro

>Here it is. Perhaps the ridiculously high OPTTOL is causing the problem; however, I suspect that it's more general than that, especially because I used the same OPTTOL successfully on the 27-atom molecule.

>Thanks.
>-Peter

This message contains the 5 kb attachment
[ corrected_input_70atoms.txt ] 70 atoms input, corrected


Message 18bfebbebrC-6769-446+00.htm, number 441, was posted on Sun Jul 13 '08 at 07:25:36
in reply to d42eec2dbHW-6768-1168+00.htm

Re^4: molecule size

Peter Grotticelli
pgrott22@optonline.net


Thanks a lot, sir; I'll try it on Monday. In return let me suggest to you to keep one of your hairs in the freezer, to preserve the DNA in the skin at the tip, for future transfer into an embryonic cell for the better regeneration of organs.

Message 9761ecf63uI-6778-1215-00.htm, number 442, was posted on Tue Jul 22 '08 at 20:15:48
MVAPICH implementation

andrea
agalparone@unict.it


Dear PC GAMESS users,
I would be interested to know what libreries have been used for compiling the MVAPICH infiniband PC GAMESS version.

Thanks in advance.

Best regards

Andrea


Message 4df668afbHW-6779-957+00.htm, number 443, was posted on Wed Jul 23 '08 at 15:57:39
in reply to 9761ecf63uI-6778-1215-00.htm

Re: MVAPICH implementation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

mvapich version of PC GAMESS is fully dynamically linked
and requires the following 32-bit libraries:

libc.so
libm.so
libpthread.so
libmpich.so
libibverbs.so

This means that mvapich should be compiled in such a way
to generate shared libs which work via ibverbs.

Regards,
Alex Granovsky



On Tue Jul 22 '08, andrea wrote
-------------------------------
>Dear PC GAMESS users,
>I would be interested to know what libreries have been used for compiling the MVAPICH infiniband PC GAMESS version.

>Thanks in advance.

>Best regards
>
>Andrea


Message 51f8f7bc3tE-6779-1380-00.htm, number 444, was posted on Wed Jul 23 '08 at 23:00:23
cannot run a job when removing internet cable!?

fabrun
fgardebien@hotmail.com


Dear users,
I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
Is it normal and is there a way for running without the internet access?
Thanks for any answer.
Best wishes.

Message 776f2952aIJ-6780-743-00.htm, number 445, was posted on Thu Jul 24 '08 at 12:24:00
error in PES run

Allan Abraham
allanbpadama@yahoo.com.ph


Greetings to all!
I am trying to view the graph of the potential energy surface of a certain structure. (Attached is the input file. ) I thought, it would work, but I encountered an error stating:
MAXIMUM GRADIENT = 31.3827498 RMS GRADIENT = 5.0271094
0OPTIMIZATION ABORTED.
-- GRADIENT OUT OF RANGE
-- MAXIMUM ALLOWED FORCE (FMAXT) = 10.000..

How would I solve this?

Thanks

Best regards,
Allan

This message contains the 3 kb attachment
[ PES.inp ] input file for PES run


Message 4df668afbHW-6782-912+00.htm, number 446, was posted on Sat Jul 26 '08 at 15:12:30
in reply to 51f8f7bc3tE-6779-1380-00.htm

Re: cannot run a job when removing internet cable!?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

This normally should not take place as PC GAMESS itself does not connect to any other computers (the only evident exception are parallel runs)

Could you please provide some more information:

1. Are you using Windows or Linux version of the PC GAMESS?
2. What is the exact PC GAMESS version and build number?
3. If running under Windows, what is the bindings dll file you are using?
4. If running under Linux, what is the MPI version PC GAMESS is linked with?
5. What is the contents of your /etc/hosts file?

Regards,
Alex Granovsky





On Wed Jul 23 '08, fabrun wrote
-------------------------------
>Dear users,
>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>Is it normal and is there a way for running without the internet access?
>Thanks for any answer.
>Best wishes.


Message 29d05261R1i-6782-1141-00.htm, number 447, was posted on Sat Jul 26 '08 at 19:02:09
how ican get the password number for this program

rdwan-alnajar
rdwan1001@maktoob.com


can any one help me to get the password

Message 4f098782ZML-6783-654+00.htm, number 448, was posted on Sun Jul 27 '08 at 10:55:40
in reply to 51f8f7bc3tE-6779-1380-00.htm

Re: cannot run a job when removing internet cable!?

Maurizio Ciofalo
omega@unipa.it


I have experienced a similar problem.
Whenever i run a distributed job in a Windows cluster (through mpiexec interface) with the Internet connection off, I get the following error:

Connection to 80.68.192.113 failed
Could not connect to lower node:
A non-recoverable error occurred during a database lookup.
(11003)

No problems at all if I am connected to the Web. I eventually attributed it to sort of spyware within NT-MPICH and resigned to.
best regards
Maurizio

On Wed Jul 23 '08, fabrun wrote
-------------------------------
>Dear users,
>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>Is it normal and is there a way for running without the internet access?
>Thanks for any answer.
>Best wishes.


Message 5bc047d55e3-6784-214-00.htm, number 449, was posted on Mon Jul 28 '08 at 03:35:29
LUMO lower in energy than HOMO

Alexey A. Popov
alxy.popov@gmail.com


Dear colleagues,

I have a problem running DFT (B3LYP) calculations of La-containing molecules. As I need a full electron density in the result of calculations (for AIM analysis), I am using UGBS basis set for La. Convergence is rather poor, but with the use of fshift parameter it is usually possible to converge SCF. However, in the result of calculations, it turns out that several occupied orbitals are usually higher in energy than some vacant orbitals. In several cases it could be cured by running one more calculation using the vectors as initial guess and reordering these orbitals (norder=1 etc in $SCF). However, I have one molecule which simply gives up to be "cured" and, however I try, it always has its HOMO higher in energy than LUMO. I would appreciate any advice on how to solve this problem.

Best regards,
Alexey Popov

PS If anybody wants to play with it, input file is attached.

This message contains the 7 kb attachment
[ La2C72-opt_ubgs.inp ] input file for PC GAMESS


Message 18bfebbebrC-6784-369-00.htm, number 450, was posted on Mon Jul 28 '08 at 06:08:47
how to make a DICTNRY (.F10) file

Peter Grotticelli
pgrott22@optonline.net


I can get .dat, .irc and .F05 outputs, but I do not get any such as .F06 through .F10, though they should appear in the scr directory. Can anybody who has used such files show me how they are made? Thanks.
- Peter

Message 18bfebbebrC-6786-1134+00.htm, number 451, was posted on Wed Jul 30 '08 at 18:54:52
in reply to 776f2952aIJ-6780-743-00.htm

Re: error in PES run

Peter Grotticelli
pgrott22@optonline.net


On Thu Jul 24 '08, Allan Abraham wrote
--------------------------------------
>Greetings to all!
>I am trying to view the graph of the potential energy surface of a certain structure. (Attached is the input file. ) I thought, it would work, but I encountered an error stating:
> � � � � �MAXIMUM GRADIENT = 31.3827498 � �RMS GRADIENT = 5.0271094
>0OPTIMIZATION ABORTED.
> � �-- GRADIENT OUT OF RANGE
> � �-- MAXIMUM ALLOWED FORCE (FMAXT) = � 10.000..

>How would I solve this?

>Thanks

>Best regards,
>Allan

Your coordinate input is in the wrong format. See the correct format under the manuals tab at http://classic.chem.msu.su/gran/gamess/index.html . I suppose you used the output from another program. You have three choices. You can manually convert; you can use a converter like OpenBabel which is found at sourceforge.net - hold shift or control, and keep the button down as long as possible, to convert multiple molecules at once; or you can write a program yourself. You need to write a program yourself to quickly convert a database with many folders in it, but if you're just working with one folder, you can simply write a program that puts the tandem output of OpenBabel into individual .inp files.

Regards to you as well,
Peter


Message 4df668afbHW-6786-1296+00.htm, number 452, was posted on Wed Jul 30 '08 at 21:36:13
in reply to 18bfebbebrC-6784-369-00.htm

Re: how to make a DICTNRY (.F10) file

Alex Granovsky
gran@classic.chem.msu.su


Hi,

PC GAMESS leaves all its binary files in its working
directory by default, unless you are using -r switch,
which means to remove them at the end of the job.

However, you need to know the internal format of
dictnry file to use it. Alternatively, you can use
irest=1 to restart from old dictnry file.

Regards,
Alex Granovsky



On Mon Jul 28 '08, Peter Grotticelli wrote
------------------------------------------
>I can get .dat, .irc and .F05 outputs, but I do not get any such as .F06 through .F10, though they should appear in the scr directory. Can anybody who has used such files show me how they are made? Thanks.
>- Peter


Message 4df668afbHW-6786-1301+00.htm, number 453, was posted on Wed Jul 30 '08 at 21:41:41
in reply to 29d05261R1i-6782-1141-00.htm

Re: how ican get the password number for this program

Alex Granovsky
gran@classic.chem.msu.su


Hi,

1. go to:

http://classic.chem.msu.su/gran/gamess/downloads.html

and follow "Registration link"

Register at ISU and explicitly select in the list PC GAMESS binaries.

You'll get the letter with the first password.

2. go to:

http://classic.chem.msu.su/gran/gamess/downloads.html

and download the desired archive.

Unzip it using first password, then run the extracted file.

Read the message, and send me all the requested information.

3. You'll get the e-mail containing second password. Use it to install PC GAMESS.

Regards,
Alex Granovsky



On Sat Jul 26 '08, rdwan-alnajar wrote
--------------------------------------
>can any one help me to get the password


Message 4df668afbHW-6786-1308+00.htm, number 454, was posted on Wed Jul 30 '08 at 21:47:40
in reply to 4f098782ZML-6783-654+00.htm

Re^2: cannot run a job when removing internet cable!?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the Winsock error number 11003 is related with problems with
host name resolution via DNS. I do not think NT-MPICH contains any spyware, as I used it many times on private cluster not connected
to Internet but having properly configured DNS server. Most likely,
you have problems with DNS on your cluster or improper "etc/hosts"
file.

Regards,
Alex



On Sun Jul 27 '08, Maurizio Ciofalo wrote
-----------------------------------------
>I have experienced a similar problem.
>Whenever i run a distributed job in a Windows cluster (through mpiexec interface) with the Internet connection off, I get the following error:

>Connection to 80.68.192.113 failed
>Could not connect to lower node:
>A non-recoverable error occurred during a database lookup.
> (11003)

>No problems at all if I am connected to the Web. I eventually attributed it to sort of spyware within NT-MPICH and resigned to.
>best regards
>Maurizio

>On Wed Jul 23 '08, fabrun wrote
>-------------------------------
>>Dear users,
>>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>>Is it normal and is there a way for running without the internet access?
>>Thanks for any answer.
>>Best wishes.


Message 4df668afbHW-6786-1308+00.htm, number 454, was edited on Wed Jul 30 '08 at 21:48:20
Re^2: cannot run a job when removing internet cable!?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the Winsock error number 11003 is related with problems with
host name resolution via DNS. I do not think NT-MPICH contains any
spyware, as I used it many times on private cluster not connected
to Internet but having properly configured DNS server. Most likely,
you have problems with DNS on your cluster or improper "etc/hosts"
file.

Regards,
Alex



On Sun Jul 27 '08, Maurizio Ciofalo wrote
-----------------------------------------
>I have experienced a similar problem.
>Whenever i run a distributed job in a Windows cluster (through mpiexec interface) with the Internet connection off, I get the following error:

>Connection to 80.68.192.113 failed
>Could not connect to lower node:
>A non-recoverable error occurred during a database lookup.
> (11003)

>No problems at all if I am connected to the Web. I eventually attributed it to sort of spyware within NT-MPICH and resigned to.
>best regards
>Maurizio

>On Wed Jul 23 '08, fabrun wrote
>-------------------------------
>>Dear users,
>>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>>Is it normal and is there a way for running without the internet access?
>>Thanks for any answer.
>>Best wishes.

[ This message was edited on Wed Jul 30 '08 by the author ]


Message 4df668afbHW-6786-1311+00.htm, number 455, was posted on Wed Jul 30 '08 at 21:50:52
in reply to c9524c90a6T-6767-1074-00.htm

Re: PCM calculate

Alex Granovsky
gran@classic.chem.msu.su


Hi,

output shows volume and area of each PCM sphere,
and at the end, the overall volume and area.

Regards,
Alex Granovsky

On Fri Jul 11 '08, Rilvia wrote
-------------------------------
>I'm using this input to calculate volume and area:

>$CONTRL UNITS=ANGS COORD=CART RUNTYP=ENERGY SCFTYP=RHF
> ICHARG= 0 MULT= 1 $END
> $GUESS GUESS= HUCKEL $END
> $SYSTEM memory=12000000 $END
> $BASIS GBASIS=N31 NGAUSS=6 POLAR=POPLE $END
> $PCM SOLVNT=WATER ICAV=1 IDISP=1 $END

> $DATA

>But in the output show me many volume and area. I would like to know what the diferences among their.

>Thaks
>Rilvia


Message d9afba4d3tE-6787-1118+00.htm, number 456, was posted on Thu Jul 31 '08 at 18:37:51
in reply to 4df668afbHW-6782-912+00.htm

Re^2: cannot run a job when removing internet cable!?

fabrun
fgardebien@hotmail.com


Dear Sir,
Sorry for my late reply and thank you for your message.
I have inserted my answers in your original text.
Thanks a lot.
Best regards.

On Sat Jul 26 '08, Alex Granovsky wrote
---------------------------------------
>Hi,

>This normally should not take place as PC GAMESS itself does not connect to any other computers (the only evident exception are parallel runs)

>Could you please provide some more information:

>1. Are you using Windows or Linux version of the PC GAMESS?
It's a linux version

>2. What is the exact PC GAMESS version and build number?
pcgamess_linux_mpich_p5
version : 6 june 1999 version 7.1 build number 4471

>3. If running under Windows, what is the bindings dll file you are using?
>4. If running under Linux, what is the MPI version PC GAMESS is linked with?
I don't really understand : maybe the information is uncluded in
the name of the package : "pcgamess_linux_mpich_p5" (?, sorry about
that)

>5. What is the contents of your /etc/hosts file?
Here it is :
----beginning of the file----
127.0.0.1 localhost me-


# The following lines are desirable for IPv6 capable hosts
# (added automatically by netbase upgrade)

::1 ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
ff02::3 ip6-allhosts

---- end THANK YOU!!

>Regards,
>Alex Granovsky
>
>
>
>
>
>On Wed Jul 23 '08, fabrun wrote
>-------------------------------
>>Dear users,
>>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>>Is it normal and is there a way for running without the internet access?
>>Thanks for any answer.
>>Best wishes.


Message 4df668afbHW-6787-1151+00.htm, number 457, was posted on Thu Jul 31 '08 at 19:11:53
in reply to 776f2952aIJ-6780-743-00.htm

Re: error in PES run

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your input file is almost correct. The main problems are:

1. You forgot to fix coordinate to scan along in the $statpt group
2. Your Z-matrix is not well-defined causing weird geometry distortions.

I'm attaching to this message archive with two sample input files.
The first one is (corrected) your initial input.
The second one (recommended) shows how to perform the same scan in DLCs

The most recent versions of PC GAMESS (available upon request) add support for bond anges and torsions scans using DLCs.

Regards,
Alex Granovsky






On Thu Jul 24 '08, Allan Abraham wrote
--------------------------------------
>Greetings to all!
>I am trying to view the graph of the potential energy surface of a certain structure. (Attached is the input file. ) I thought, it would work, but I encountered an error stating:
> MAXIMUM GRADIENT = 31.3827498 RMS GRADIENT = 5.0271094
>0OPTIMIZATION ABORTED.
> -- GRADIENT OUT OF RANGE
> -- MAXIMUM ALLOWED FORCE (FMAXT) = 10.000..

>How would I solve this?

>Thanks

>Best regards,
>Allan

This message contains the 2 kb attachment
[ rsurf_samples.rar ] Samples for rsurface runs


Message d4c0f196bHW-6791-843+00.htm, number 458, was posted on Mon Aug 4 '08 at 14:03:06
in reply to d9afba4d3tE-6787-1118+00.htm

Re^3: cannot run a job when removing internet cable!?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

assume that your IP address is 192.168.1.48
and host name is phys0bha-18.physclass.msu.ru

Then, you would need to add the following to your /etc/hosts file

127.0.0.1 localhost me
---beginning of changes---
192.168.1.48 phys0bha-18 phys0bha-18.physclass.msu.ru
---end of changes---

This should help, if it does not, try to use -nompi PC GAMESS command line option - it will solve your problems but you'll be limited to serial jobs only.

Regards,
Alex Granovsky


On Thu Jul 31 '08, fabrun wrote
-------------------------------
>Dear Sir,
>Sorry for my late reply and thank you for your message.
>I have inserted my answers in your original text.
>Thanks a lot.
>Best regards.

>On Sat Jul 26 '08, Alex Granovsky wrote
>---------------------------------------
>>Hi,

>>This normally should not take place as PC GAMESS itself does not connect to any other computers (the only evident exception are parallel runs)

>>Could you please provide some more information:

>>1. Are you using Windows or Linux version of the PC GAMESS?
>It's a linux version

>>2. What is the exact PC GAMESS version and build number?
>pcgamess_linux_mpich_p5
>version : 6 june 1999 version 7.1 build number 4471

>>3. If running under Windows, what is the bindings dll file you are using?
>>4. If running under Linux, what is the MPI version PC GAMESS is linked with?
>I don't really understand : maybe the information is uncluded in
>the name of the package : "pcgamess_linux_mpich_p5" (?, sorry about
>that)

>>5. What is the contents of your /etc/hosts file?
>Here it is :
>----beginning of the file----
>127.0.0.1 localhost me-
>
>
># The following lines are desirable for IPv6 capable hosts
># (added automatically by netbase upgrade)

>::1 ip6-localhost ip6-loopback
>fe00::0 ip6-localnet
>ff00::0 ip6-mcastprefix
>ff02::1 ip6-allnodes
>ff02::2 ip6-allrouters
>ff02::3 ip6-allhosts

>---- end THANK YOU!!

>>Regards,
>>Alex Granovsky
>>
>>
>>
>>
>>
>>On Wed Jul 23 '08, fabrun wrote
>>-------------------------------
>>>Dear users,
>>>I am very new in the PCGamess world, so this one may be naive: I cannot run a job when I remove the internet cable from me PC. There are no errors, it just take the input but nothing happens at all and when I plug the cable back, it runs perfectly.
>>>Is it normal and is there a way for running without the internet access?
>>>Thanks for any answer.
>>>Best wishes.


Message d4c0f196bHW-6797-1297+00.htm, number 459, was posted on Sun Aug 10 '08 at 21:42:53
in reply to 5bc047d55e3-6784-214-00.htm

Re: LUMO lower in energy than HOMO

Alex Granovsky
gran@classic.chem.msu.su


Dear Lesha,

First, I began with fixing old PC GAMESS bug which did not allow
Huckel guess for La with all-electron basis sets.

Second, to save time I used somewhat different but similar basis
sets on Carbons and quite a different basis set on La (midi).

Frankly, I do not like your choice of basis set on La. While it is
fully uncontracted which assumes high flexibility, it does not
contain any polarization functions (i.e., f and g harmonics).
I do not think it is suitable for your purposes. According to my
calculations (below), the f-functions are important, and maybe
g-functions would be of some use as well. This is because of
non-trivial configuration of bonds formed between La and C.
The improper choice of basis set can be one of the reasons
of convergence to the improper SCF solution in your case
(another could be Hcore guess).

After selecting basis sets, I ran RHF calculations with default
settings, and used RHF orbitals as the starting guess for subsequent
DFT. Both calculations completed without any problems and resulted in
the correct ordering of orbital energies.

I'm attaching two inputs and outputs to this post and hope
they could be of some help for you.

Regards,


On Mon Jul 28 '08, Alexey A. Popov wrote
----------------------------------------
>Dear colleagues,

>I have a problem running DFT (B3LYP) calculations of La-containing molecules. As I need a full electron density in the result of calculations (for AIM analysis), I am using UGBS basis set for La. Convergence is rather poor, but with the use of fshift parameter it is usually possible to converge SCF. However, in the result of calculations, it turns out that several occupied orbitals are usually higher in energy than some vacant orbitals. In several cases it could be cured by running one more calculation using the vectors as initial guess and reordering these orbitals (norder=1 etc in $SCF). However, I have one molecule which simply gives up to be "cured" and, however I try, it always has its HOMO higher in energy than LUMO. I would appreciate any advice on how to solve this problem.

>Best regards,
>Alexey Popov

>PS If anybody wants to play with it, input file is attached.

This message contains the 11470 kb attachment
[ lesha.rar ] Inputs and outputs for Alexey's problem


Message d4c0f196bHW-6797-1305+00.htm, number 459, was edited on Sun Aug 10 '08 at 21:45:17
and replaces message d4c0f196bHW-6797-1297+00.htm

Re: LUMO lower in energy than HOMO

Alex Granovsky
gran@classic.chem.msu.su


Dear Lesha,

First, I began with fixing old PC GAMESS bug which did not allow
Huckel guess for La with all-electron basis sets. Please let me
know if you need these binaries.

Second, to save time I used somewhat different but similar basis
sets on Carbons and quite a different basis set on La (namely, MIDI).

Frankly, I do not like your choice of basis set on La. While it is
fully uncontracted which assumes high flexibility, it does not
contain any polarization functions (i.e., f and g harmonics).
I do not think it is suitable for your purposes. According to my
calculations (below), the f-functions are important, and maybe
g-functions would be of some help as well. This is because of
non-trivial configuration of bonds formed between La and C.
The improper choice of basis set can be one of the reasons
of convergence to the wrong SCF solution in your case
(another could be the Hcore guess).

After selecting basis sets, I ran RHF calculations with default
settings, and used RHF orbitals as the starting guess for the
subsequent DFT. Both calculations completed without any problems
and resulted in the correct ordering of orbital energies.

I'm attaching two inputs and outputs to this post and hope
they could be of some help for you.

Regards,
Alex Granovsky

On Mon Jul 28 '08, Alexey A. Popov wrote
----------------------------------------
>Dear colleagues,

>I have a problem running DFT (B3LYP) calculations of La-containing molecules. As I need a full electron density in the result of calculations (for AIM analysis), I am using UGBS basis set for La. Convergence is rather poor, but with the use of fshift parameter it is usually possible to converge SCF. However, in the result of calculations, it turns out that several occupied orbitals are usually higher in energy than some vacant orbitals. In several cases it could be cured by running one more calculation using the vectors as initial guess and reordering these orbitals (norder=1 etc in $SCF). However, I have one molecule which simply gives up to be "cured" and, however I try, it always has its HOMO higher in energy than LUMO. I would appreciate any advice on how to solve this problem.

>Best regards,
>Alexey Popov

>PS If anybody wants to play with it, input file is attached.

This message contains the 11470 kb attachment
[ lesha.rar ] Inputs and outputs for Alexey's problem

[ This message was edited on Sun Aug 10 '08 by the author ]


Message 50da3cbd0Mg-6799-719-00.htm, number 460, was posted on Tue Aug 12 '08 at 11:58:42
Memory issue on parallel run,

Daniele Passerone
daniele.passerone@empa.ch


Dear experts,

This is my first message in this forum.
I am running PCGAMESS on a linux cluster with dual core Woodcrest processors (4 cores per node, 8 GB RAM per node) and gigabit interconnect.

I have - I suppose - memory issues on a particular job, B3LYP with 152 atoms, and a total number of basis functions (6-311(d,p)) is about 2400.
The EXETYP=CHECK of my jobs works, but when I really run the SCF, at a certain point (depending on the memory requested, and the number of nodes) the jobs "freezes" and each node shows only 3 processes instead of four: the 4th one remains in "S"leeping state.
Sometimes this happens during the initial guess, sometimes after the first SCF iteration, sometimes after a couple...
After a while the job stops for good, without any particular message, even when I run pcgamess with the -V option.

Here the first lines of my

$CONTRL SCFTYP=RHF RUNTYP=ENERGY COORD=UNIQUE
DFTTYP=B3LYP1 EXETYP=CHECK MAXIT=200 MULT=1
$END
$SYSTEM MEMORY=250000000 $END
$BASIS NDFUNC=1 NPFUNC=1 GBASIS=N311 NGAUSS=6 $END
$GUESS GUESS=HUCKEL $END
$SCF DIRSCF=.TRUE. $END
$DFT NRAD=75 $END
$DATA
Model1
Cn 2

O 8.0 10.0219254515 6.5718610530 0.0000000000
O 8.0 10.0595821413 4.3540900000 0.0000000000
...



and my processor file:

local 3
node003 4 /home/apps/bin/pcgamess_d
node007 4 /home/apps/bin/pcgamess_d
node008 4 /home/apps/bin/pcgamess_d
node009 4 /home/apps/bin/pcgamess_d
node010 4 /home/apps/bin/pcgamess_d


Any help would be really appreciated.

Best,
Daniele Passerone,
Empa, Zrich


Message 50da3cbd0Mg-6801-1393+00.htm, number 461, was posted on Thu Aug 14 '08 at 23:12:59
in reply to 50da3cbd0Mg-6799-719-00.htm

Re: Memory issue on parallel run,

Daniele Passerone
daniele.passerone@empa.ch


On Tue Aug 12 '08, Daniele Passerone wrote
------------------------------------------
>Dear experts,

>After a while the job stops for good, without any particular message, even when I run pcgamess with the -V option.
>

After 2-3 days (and much more weeks before this post) of continuous email exchange with Prof. Granovsky, my PCGAMESS works normally: thanks Alex!
The point is that our cluster works with a particular communication protocol, a flavor of mpich that bypasses tcp/ip on the gigabit. This is called "parastation" and comes from the company par-tec (http://www.par-tec.com).
While Alex Granovsky was trying to compile pcgamess with the parastation mpi libraries,
I was trying to use the standard statically linked version of pcgamess, which worked on
single node and gave the abovementioned problem on multiple nodes.
According to Alex, for large jobs

AG>MPICH uses communications via shared memory while running on single node,
AG>and TCP/IP to communicate between nodes. It seems you have some problems
AG>with intensive communication over TCP/IP on your cluster. This could be due to
AG>mpich bugs, or (most likely) because of some conflict with parastation software.

Indeed, today I received the latest version and I can say that the problem was solved
and the issue closed

AG>Thus I believe the origin of your problems was incompatibility
AG>between mpich and parastation-configured cluster.

I will soon post results on standard benchmarks.

Yours,

Daniele
>


Message 50da3cbd0Mg-6802-58+00.htm, number 461, was edited on Fri Aug 15 '08 at 00:57:42
and replaces message 50da3cbd0Mg-6801-1393+00.htm

Re: Memory issue on parallel run,

Daniele Passerone
daniele.passerone@empa.ch


On Tue Aug 12 '08, Daniele Passerone wrote
------------------------------------------
>Dear experts,

>After a while the job stops for good, without any particular message, even when I run pcgamess with the -V option.
>

After 2-3 days (and much more weeks before this post) of continuous email exchange with Prof. Granovsky, my PCGAMESS works normally: thanks Alex!
The point is that our cluster works with a particular communication protocol, a flavor of mpich that bypasses tcp/ip on the gigabit. This is called "parastation" and comes from the company par-tec (http://www.par-tec.comhttp://www.par-tec.com).
While Alex Granovsky was trying to compile pcgamess with the parastation mpi libraries,
I was trying to use the standard mpich-dynamically linked version of pcgamess, which worked on
single node and gave the abovementioned problem on multiple nodes.
According to Alex, for large jobs

AG>MPICH uses communications via shared memory while running on single node,
AG>and TCP/IP to communicate between nodes. It seems you have some problems
AG>with intensive communication over TCP/IP on your cluster. This could be due to
AG>mpich bugs, or (most likely) because of some conflict with parastation software.

Indeed, today I received the latest version and I can say that the problem was solved
and the issue closed

AG>Thus I believe the origin of your problems was incompatibility
AG>between mpich and parastation-configured cluster.

I will soon post results on standard benchmarks.

Yours,

Daniele
>

[ This message was edited on Fri Aug 15 '08 by the author ]


Message 0ccb39185au-6803-1191-00.htm, number 462, was posted on Sat Aug 16 '08 at 19:51:41
coumpound jobs

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


Hi,

two questions.

1. the biggest thing GAMESS in general has been lacking has been an ability to run compound jobs from one input, e.g. low level optimization, followed by high level, followed by SP etc. Almost all other softwares have it, why don't GAMESS? Maybe that could be implemented in PCGAMESS? It's such a painful old fashioned way copying $VEC for a guess...

2. I have tons of memory but when job is running is being utilized 5% (uhf optimization). Is there a way to crank the memory usage up?

thanks

Jonas


Message d4c0fb6f2GM-6804-838+00.htm, number 463, was posted on Sun Aug 17 '08 at 13:59:24
in reply to 0ccb39185au-6803-1191-00.htm

Re: coumpound jobs

Igor Polyakov
polyakoviv@gmail.com


Hello!

>1. the biggest thing GAMESS in general has been lacking has been an ability to run compound jobs from one input, e.g. low level optimization, followed by high level, followed by SP etc. Almost all other softwares have it, why don't GAMESS? Maybe that could be implemented in PCGAMESS? It's such a painful old fashioned way copying $VEC for a guess...

There exists a macrolanguage for PC GAMESS, U can download it from this website. I suppose everyone are encouraged to write additional software/scripts for automatization of calculations.

>2. I have tons of memory but when job is running is being utilized 5% (uhf optimization). Is there a way to crank the memory usage up?

UHF optimization does not require vast amounts of memory in case if U do not use several thousands of basis functions in ur calculations. MP2 and MRMP2 consume a lot of memory and even more memory if ur tasks use hundreds of effective fragments. Also I should remind about $system MWORDS parameter which defines maximum amount of memory for PC GAMESS process. Choose it wisely- first time on each job set it to maximum for ur computer, then run exetyp=check and see what exact amount of memory will be used by PC GAMESS and reset MWORDS to somewhat above this value. There is a way to determine ur computer's MWORDS maximum. If U have 2G memory and u run 1 instance of PC GAMESS u should do the following calculation (2048-128)/8=240=MWORDS, this should work on Linux but not windows. As for my experience maximum u can get properly working on WinXP is 200-220MWORDS from 2G of memory. If u run PC GAMESS in MPI (e.g. multiple instances) u should divide these 200mwords equally amongst all instances, e.g. if 2, then 100, if 4 then 50. This is relevant for multiprocessor or multicore system or both=).
Will be glad if my answer could help U.

Best luck, Igor.


Message 0ccb39185au-6805-244-00.htm, number 464, was posted on Mon Aug 18 '08 at 04:03:49
some keywords don't work

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


Hi,

is there a way to lift SCF convergence from 2x10-5 up? I have particularly tough to converge systems and trying to preoptimize with MINI. Problem is that it's oscillating near 1x10-4 and not converging.

That's odd, but changing keyword CONV has not been recognized and simply aborted calculation as if the keyword was illegal. So did trying to change NOCONV to true. It's aborting as if this keyword was illegal. Have some of the GAMESS keywords been changed in PCGAMESS?


Message d4c0fb6f2GM-6805-709+00.htm, number 465, was posted on Mon Aug 18 '08 at 11:48:52
in reply to 0ccb39185au-6805-244-00.htm

Re: some keywords don't work

Igor Polyakov
polyakoviv@gmail.com


Hello,

>is there a way to lift SCF convergence from 2x10-5 up? I have particularly tough to converge systems and trying to preoptimize with MINI. Problem is that it's oscillating near 1x10-4 and not converging.

In manual it is said:

NCONV = n SCF density convergence criteria.
Convergence is reached when the density change
between two consecutive SCF cycles is less than
10.0**(-n) in absolute value. One more cycle
is executed after reaching convergence. Less
accuracy in NCONV gives questionable gradients.
(default is n=5, except CI or MP2 gradients n=6)

And this one really works- i've just checked. So add $scf nconv=4 $end In case of difficult convergence u can also try to change the convergers ( switch to DIIS from SOSCF for example). Also sometimes u can converge the case with higher multiplicity and use this guess as a start for ur original calculation or if u use HF u can compute DFT (as from my experience DFT converges better) orbitals and use them as a guess. There is also a lot of trickier ways.

>That's odd, but changing keyword CONV has not been recognized and simply aborted calculation as if the keyword was illegal. So did trying to change NOCONV to true. It's aborting as if this keyword was illegal. Have some of the GAMESS keywords been changed in PCGAMESS?

there is no such option NOCONV for PC GAMESS - and also it is not described in manual. Please use input.doc from default PC GAMESS distribution and for some specific options- manuals from "readmes" directory. Also on this webpage u can find some additional manuals.

Best luck, Igor


Message 0ccb39185au-6805-961+00.htm, number 466, was posted on Mon Aug 18 '08 at 16:01:12
in reply to d4c0fb6f2GM-6805-709+00.htm

Re^2: some keywords don't work

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


Tried changing NCONV, but now I just can't submit the job

# /root/pcgamess/pcgamess -r -f -p -i /root/pcgamess/input -o /root/pcgamess/output.out -ex /root/pcgamess -t /root/pcgamess/task123
p0_18746: p4_error: semget failed for setnum: 0


On Mon Aug 18 '08, Igor Polyakov wrote
--------------------------------------
>Hello,

>>is there a way to lift SCF convergence from 2x10-5 up? I have particularly tough to converge systems and trying to preoptimize with MINI. Problem is that it's oscillating near 1x10-4 and not converging.

>In manual it is said:

> NCONV = n SCF density convergence criteria.
> Convergence is reached when the density change
> between two consecutive SCF cycles is less than
> 10.0**(-n) in absolute value. One more cycle
> is executed after reaching convergence. Less
> accuracy in NCONV gives questionable gradients.
> (default is n=5, except CI or MP2 gradients n=6)

>And this one really works- i've just checked. So add $scf nconv=4 $end In case of difficult convergence u can also try to change the convergers ( switch to DIIS from SOSCF for example). Also sometimes u can converge the case with higher multiplicity and use this guess as a start for ur original calculation or if u use HF u can compute DFT (as from my experience DFT converges better) orbitals and use them as a guess. There is also a lot of trickier ways.

>>That's odd, but changing keyword CONV has not been recognized and simply aborted calculation as if the keyword was illegal. So did trying to change NOCONV to true. It's aborting as if this keyword was illegal. Have some of the GAMESS keywords been changed in PCGAMESS?

>there is no such option NOCONV for PC GAMESS - and also it is not described in manual. Please use input.doc from default PC GAMESS distribution and for some specific options- manuals from "readmes" directory. Also on this webpage u can find some additional manuals.

>Best luck, Igor


Message d4c0fb6f2GM-6805-1030+00.htm, number 467, was posted on Mon Aug 18 '08 at 17:10:21
in reply to 0ccb39185au-6805-961+00.htm

Re^3: some keywords don't work

Igor Polyakov
polyakoviv@gmail.com


Hello!

>Tried changing NCONV, but now I just can't submit the job

># /root/pcgamess/pcgamess -r -f -p -i /root/pcgamess/input -o /root/pcgamess/output.out -ex /root/pcgamess -t /root/pcgamess/task123
>p0_18746: p4_error: semget failed for setnum: 0

This error has nothing to do with NCONV or even PC GAMESS itself. This error means that you need to clean shared memory segments and semaphore arrays and is caused by MPICH. First of all for pure sequential execution u can use -nompi key in the very end of the PC GAMESS commandline. Secondly, in the MPICH distribution there is a script called 'cleanipcs' which cleans shared memory segments and semaphore arrays. Simply make a call to it. To see the 'status' of ur linux shared memory type 'ipcs' command -u will see the list.

Best luck, Igor


Message 54d942e6a7C-6815-920-00.htm, number 468, was posted on Thu Aug 28 '08 at 15:20:29
Good basic book on calculation?

jaxz
jaxzjaxz@gmail.com


Dear all,
can you please recommend a good book that introduces the reader in a rather easy way to the subject of ab initio calculations.

I may take on a student to do some basic calculations on caffeine-like molecules to determine vibrational frequencies and then check these with actual spectroscopic experiments. This is a rather brief project spanning 12-16 weeks.

What litterature is a good way to start?

Best regards,
jaxz
dept of physics
umea university
Sweden


Message c6b48310SZP-6816-1244+00.htm, number 469, was posted on Fri Aug 29 '08 at 20:45:38
in reply to 54d942e6a7C-6815-920-00.htm

Re: Good basic book on calculation?

Slawomir Janicki
slawomir.janicki@comcast.net


Hi jaxz

The three books I found most useful are:

"Essentials of Computational Chemistry: Theories and Models by Christopher" J. Cramer
"Introduction to Computational Chemistry" by Frank Jensen and
"Computational Organic Chemistry" by Steven M. Bachrach

Given the short project time I would actually avoid throwing a book at the student. I assume that the student has a basic knowledge of quantum chemistry, so I would suggest one of many on-line tutorials on how to run GAMESS/PC-GAMESS. Let him/her start with actual building the caffeine model (or something smaller) input file, even if the initial set of options would not be optimal. Then, work with the student to optimize this calculation and transfer the experience to the actual system.

At the end of the project your student probably would not know everything about PC-GAMESS, but should be able to correctly put together a basic optimization and hessian job. This would be a very solid basis for future progress.

HTH

Slawomir Janicki

On Thu Aug 28 '08, jaxz wrote
-----------------------------
>Dear all,
>can you please recommend a good book that introduces the reader in a rather easy way to the subject of ab initio calculations.

>I may take on a student to do some basic calculations on caffeine-like molecules to determine vibrational frequencies and then check these with actual spectroscopic experiments. This is a rather brief project spanning 12-16 weeks.

>What litterature is a good way to start?

>Best regards,
>jaxz
>dept of physics
>umea university
>Sweden


Message 511c224226n-6818-829-00.htm, number 470, was posted on Sun Aug 31 '08 at 13:49:43
Restart in geometry optimization and hessian calculation

Jamshid
j.najafpour@gmail.com


Dear Sir:

How can I restart geometry optimization and hessian calculation when computing process have stoped.

Best Regards
Jamshid


Message 4245d5eaNOu-6818-1142-00.htm, number 471, was posted on Sun Aug 31 '08 at 19:05:06
Why are the O3LYP5 energies of PC-GAMESS/ORCA/G03/PQS so different?

Wenli Zou
zouwl@mail.utexas.edu


Following are the O3LYP(VWN5)/cc-pVTZ(spherical) energies of F-,

-99.698059008 pc-gamess
-99.698015916 orca
-99.811519454 gaussian03
-100.262303658 pqs

We can see that the energies are very different (except PC-GAMESS and ORCA). If other functionals are used (OLYP, B3LYP, and so on), the results are basically the same, so the large energy difference can not be resulted by the basis set, grid scheme, and precision.

Best wishes,

Wenli

By the way, NWChem can also do O3LYP calculation by defining

dft
XC vwn_5 0.19 lyp 0.81 Hfexch 0.1161 slater 0.9262 OPTX 0.8133
end

However, if OPTX is used, the coefficient of Slater will be reset to 1.05151 (for reasons see dftxca.src of GAMESS-US). So the energy of NWChem (-100.885076297) must be wrong.


Message 4df668afbHW-6818-1335+00.htm, number 472, was posted on Sun Aug 31 '08 at 22:14:47
in reply to 4245d5eaNOu-6818-1142-00.htm

Re: Why are the O3LYP5 energies of PC-GAMESS/ORCA/G03/PQS so different?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

First, there are both O3LYP1 (O3LYP/VWN1) and O3LYP5 (O3LYP/VWN5)
functionals available in the PC GAMESS. The O3LYP5 is the default,
with O3LYP being simply the shortcut to O3LYP5.

Second, I can confirm that O3LYP5 in PC GAMESS can be described
in the NWChem terms exactly as:

dft
XC vwn_5 0.19 lyp 0.81 Hfexch 0.1161 slater 0.9262 OPTX 0.8133
end

The only exception is that PC GAMESS (unlike NWChem) uses
correct multiplier for Slater's contribution.

Finally, it seems to me that PC GAMESS and ORCA produces the same
results within the precision of the particular integration scheme.
Thus, most likely Gaussian's and pqs's numbers are wrong.

Regards,
Alex Granovsky



On Sun Aug 31 '08, Wenli Zou wrote
----------------------------------
>Following are the O3LYP(VWN5)/cc-pVTZ(spherical) energies of F-,

> -99.698059008 pc-gamess
> -99.698015916 orca
> -99.811519454 gaussian03
>-100.262303658 pqs

>We can see that the energies are very different (except PC-GAMESS and ORCA). If other functionals are used (OLYP, B3LYP, and so on), the results are basically the same, so the large energy difference can not be resulted by the basis set, grid scheme, and precision.

>Best wishes,

>Wenli

>By the way, NWChem can also do O3LYP calculation by defining

>dft
>XC vwn_5 0.19 lyp 0.81 Hfexch 0.1161 slater 0.9262 OPTX 0.8133
>end

>However, if OPTX is used, the coefficient of Slater will be reset to 1.05151 (for reasons see dftxca.src of GAMESS-US). So the energy of NWChem (-100.885076297) must be wrong.
>


Message 511c224226n-6819-750-00.htm, number 473, was posted on Mon Sep 1 '08 at 12:30:13
Restart in geometry optimization and hessian calculations

Jamshid Najafpour
j.najafpour@gmail.com


How can I restart geometry optimization and hessian calculation when the computing process have stoped.

Best Regards
Jamshid


Message d42eec0cbHW-6819-1433+00.htm, number 474, was posted on Mon Sep 1 '08 at 23:52:59
in reply to 511c224226n-6819-750-00.htm

Re: Restart in geometry optimization and hessian calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

For restart of optimization itself, just take the last geometry
from the output or punch file, replace the old one in the
input file and resubmit the job.

For restart of optimization+hessian job stopped at the stage of
numerical hessian calculation, replace initial geometry by the
final one, and add $vib group from the ircdata file to your input
(do not forget to put $end at the end). Then change runtyp to
hessian and resubmit. Optionally, you can provide converged
$vec group and $grad group from the punch file at found
stationary point as well

For restart of optimization+hessian job stopped at the stage of
analytical hessian calculation, replace initial geometry by the
final one, change runtyp to hessian and resubmit. Optionally,
you can provide $vec group from the punch file

Regards,
Alex Granovsky



On Mon Sep 1 '08, Jamshid Najafpour wrote
-----------------------------------------
>How can I restart geometry optimization and hessian calculation when the computing process have stoped.

>Best Regards
>Jamshid

>


Message 4245d5eaNOu-6820-1+00.htm, number 475, was posted on Tue Sep 2 '08 at 00:02:38
in reply to d42eec0cbHW-6819-1433+00.htm

Gaussian's O3LYP is wrong!

Wenli Zou
zouwl@mail.utexas.edu


Thank you!

I modified the source code of NWChem. Now the energy is -99.698055076796 a.u., being very close to the ones of PC-GAMESS and ORCA.

Gaussian's O3LYP is indeed wrong. To correct it, the parameters of the correlation functionals must be specified explicitly by

# cc-pvtz IOp(3/74=-24) IOP(3/77=0813309262)

or (for Gaussian 03.c.01 and higher)

# o3lyp/cc-pvtz IOP(3/77=0813309262)

By this way, a correct energy of -99.6980592818 a.u. can be obtained.

A test for F2 by O3LYP5/cc-pvtz with Grid=UltraFine.

bond distance (Re) /Ang.
1.3897 (G03's original O3LYP)
1.3788 (G03's corrected O3LYP)
1.3787 (PC-GAMESS)

energy /a.u.
-199.526380102 (G03's original O3LYP)
-199.303011565 (G03's corrected O3LYP)
-199.303013311 (PC-GAMESS)

HOMO-LUMO gap /eV
6.0 (G03's original O3LYP)
6.1 (G03's corrected O3LYP)
6.1 (PC-GAMESS)

dissociation energy (De) /eV
2.6 (G03's original O3LYP)
2.5 (G03's corrected O3LYP)
2.5 (PC-GAMESS)


Message 4245d5eaNOu-6820-6+00.htm, number 476, was posted on Tue Sep 2 '08 at 00:06:35
in reply to 4df668afbHW-6818-1335+00.htm

Gaussian's O3LYP is wrong!

Wenli Zou
zouwl@mail.utexas.edu


Thank you!

You are right. I modified the source code of NWChem. Now the energy is -99.698055076796 a.u., being very close to the ones of PC-GAMESS and ORCA.

Gaussian's O3LYP is indeed wrong. To correct it, the parameters of the correlation functionals must be specified explicitly by

# cc-pvtz IOp(3/74=-24) IOP(3/77=0813309262)

or (for Gaussian 03.c.01 and higher)

# o3lyp/cc-pvtz IOP(3/77=0813309262)

By this way, a correct energy of -99.6980592818 a.u. can be obtained.

A test for F2 by O3LYP5/cc-pvtz with Grid=UltraFine.

bond distance (Re) /Ang.
1.3897 (G03's original O3LYP)
1.3788 (G03's corrected O3LYP)
1.3787 (PC-GAMESS)

energy /a.u.
-199.526380102 (G03's original O3LYP)
-199.303011565 (G03's corrected O3LYP)
-199.303013311 (PC-GAMESS)

HOMO-LUMO gap /eV
6.0 (G03's original O3LYP)
6.1 (G03's corrected O3LYP)
6.1 (PC-GAMESS)

dissociation energy (De) /eV
2.6 (G03's original O3LYP)
2.5 (G03's corrected O3LYP)
2.5 (PC-GAMESS)


Message 55c62e6b26n-6820-1285+00.htm, number 477, was posted on Tue Sep 2 '08 at 21:25:24
in reply to d42eec0cbHW-6819-1433+00.htm

Re^2: Restart in geometry optimization and hessian calculations

Jamshid Najafpour
j.najafpour@gmail.com


Hi:

METOD=SEMINUM is the default in DFT calculation, but in B3LYP/aug-cc-pVDZ calculation I dont have IRCDATA file. How can I find ircdata file? Where is it?

Best Whishes
Jamshid
On Mon Sep 1 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>For restart of optimization itself, just take the last geometry
>from the output or punch file, replace the old one in the
>input file and resubmit the job.

>For restart of optimization+hessian job stopped at the stage of
>numerical hessian calculation, replace initial geometry by the
>final one, and add $vib group from the ircdata file to your input
>(do not forget to put $end at the end). Then change runtyp to
>hessian and resubmit. Optionally, you can provide converged
>$vec group and $grad group from the punch file at found
>stationary point as well

>For restart of optimization+hessian job stopped at the stage of
>analytical hessian calculation, replace initial geometry by the
>final one, change runtyp to hessian and resubmit. Optionally,
>you can provide $vec group from the punch file

>Regards,
>Alex Granovsky
>
>
>
>On Mon Sep 1 '08, Jamshid Najafpour wrote
>-----------------------------------------
>>How can I restart geometry optimization and hessian calculation when the computing process have stoped.

>>Best Regards
>>Jamshid

>>


Message 4df668afbHW-6821-1058+00.htm, number 478, was posted on Wed Sep 3 '08 at 17:38:15
in reply to 55c62e6b26n-6820-1285+00.htm

Re^3: Restart in geometry optimization and hessian calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

are you sure you are using PC GAMESS, not GAMESS (US)?

Regards,
Alex Granovsky


On Tue Sep 2 '08, Jamshid Najafpour wrote
-----------------------------------------
>Hi:

>METOD=SEMINUM is the default in DFT calculation, but in B3LYP/aug-cc-pVDZ calculation I dont have IRCDATA file. How can I find ircdata file? Where is it?

>Best Whishes
>Jamshid
>On Mon Sep 1 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>For restart of optimization itself, just take the last geometry
>>from the output or punch file, replace the old one in the
>>input file and resubmit the job.

>>For restart of optimization+hessian job stopped at the stage of
>>numerical hessian calculation, replace initial geometry by the
>>final one, and add $vib group from the ircdata file to your input
>>(do not forget to put $end at the end). Then change runtyp to
>>hessian and resubmit. Optionally, you can provide converged
>>$vec group and $grad group from the punch file at found
>>stationary point as well

>>For restart of optimization+hessian job stopped at the stage of
>>analytical hessian calculation, replace initial geometry by the
>>final one, change runtyp to hessian and resubmit. Optionally,
>>you can provide $vec group from the punch file

>>Regards,
>>Alex Granovsky
>>
>>
>>
>>On Mon Sep 1 '08, Jamshid Najafpour wrote
>>-----------------------------------------
>>>How can I restart geometry optimization and hessian calculation when the computing process have stoped.

>>>Best Regards
>>>Jamshid

>>>


Message 55c62e5026n-6823-9+00.htm, number 479, was posted on Fri Sep 5 '08 at 00:12:24
in reply to 4df668afbHW-6821-1058+00.htm

Re^4: Restart in geometry optimization and hessian calculations

Jamshid Najafpour
j.najafpour@gmail.com


Hi,

I found it. It was in TEMP directory.
I am using WinGAMESS program.

Wishes
Jamshid
On Wed Sep 3 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>are you sure you are using PC GAMESS, not GAMESS (US)?

>Regards,
>Alex Granovsky
>
>
>On Tue Sep 2 '08, Jamshid Najafpour wrote
>-----------------------------------------
>>Hi:

>>METOD=SEMINUM is the default in DFT calculation, but in B3LYP/aug-cc-pVDZ calculation I dont have IRCDATA file. How can I find ircdata file? Where is it?

>>Best Whishes
>>Jamshid
>>On Mon Sep 1 '08, Alex Granovsky wrote
>>--------------------------------------
>>>Hi,

>>>For restart of optimization itself, just take the last geometry
>>>from the output or punch file, replace the old one in the
>>>input file and resubmit the job.

>>>For restart of optimization+hessian job stopped at the stage of
>>>numerical hessian calculation, replace initial geometry by the
>>>final one, and add $vib group from the ircdata file to your input
>>>(do not forget to put $end at the end). Then change runtyp to
>>>hessian and resubmit. Optionally, you can provide converged
>>>$vec group and $grad group from the punch file at found
>>>stationary point as well

>>>For restart of optimization+hessian job stopped at the stage of
>>>analytical hessian calculation, replace initial geometry by the
>>>final one, change runtyp to hessian and resubmit. Optionally,
>>>you can provide $vec group from the punch file

>>>Regards,
>>>Alex Granovsky
>>>
>>>
>>>
>>>On Mon Sep 1 '08, Jamshid Najafpour wrote
>>>-----------------------------------------
>>>>How can I restart geometry optimization and hessian calculation when the computing process have stoped.

>>>>Best Regards
>>>>Jamshid

>>>>


Message d4c0fb6e2GM-6825-81+00.htm, number 480, was posted on Sun Sep 7 '08 at 01:20:40
in reply to 55c62e5026n-6823-9+00.htm

Re^5: Restart in geometry optimization and hessian calculations

Igor Polyakov
polyakoviv@gmail.com


Jamshid,

The PC GAMESS programme is kind of different from the WinGAMESS. It should be noted that PC GAMESS performs much faster in calculations than WinGAMESS (you can see it for yourself by looking the test results comparison), easily runs in parallel on Windows clusters, has some bonus options compared to WinGAMESS, so unless your calculations always do require the very specific options, that are presented only in WinGAMESS and are absent in PC GAMESS, you should think about switching from WinGAMESS to PC GAMESS.

Best luck, Igor


Message 5395f7047Tv-6828-1053-00.htm, number 481, was posted on Wed Sep 10 '08 at 17:33:32
Problem in computing thermochemistry of small molecules (e.g. CO & CO2)

Denis Zavelev
denis.zavelev@gmail.com


I'm trying to compute thermochemistry of some small molecules (namely CO, CO2 and O2) in order to estimate thermochemistry of the reaction studied.
It doesn't matter what to give in input file: non-optimized geometry, optimized geometry without $HESS or with it - I am getting the following:
=========
*** ERROR *** ATTEMPT TO READ A DAF RECORD THAT WAS NEVER WRITTEN.
RECORD NUMBER 49 OF LENGTH 0 DOES NOT EXIST.
CHECK -PROG.DOC- FOR A LIST OF DIRECT ACCESS FILE CONTENTS
=========


Here's input file (for one of the attempts):
=========
$CONTRL RUNTYP=OPTIMIZE ECP=SBKJC COORD=CART DFTTYP=B3LYP NOSYM=1
MAXIT=300 SCFTYP=RHF MULT=1 $END
$SYSTEM TIMLIM=30000 MWORDS=40 $END
$BASIS GBASIS=SBKJC $END
$SCF DAMP=.t. SHIFT=.t. NCONV=6 $END
$STATPT NSTEP=500 HSSEND=.t. OPTTOL=0.00001 $END
$ELFLDG IEFLD=2 $END
$ELDENS IEDEN=1 $END
$FORCE TEMP=453 PRTSCN=.t. DECOMP=.t. $END
$DATA
CO
C1
CARBON 6.0 -1.225552804 -2.699077973 0.189065492
OXYGEN 8.0 0.201717658 -2.610859894 0.193139549
$END
=========

What am I doing wrong?


Message 4df668afbHW-6829-1148+00.htm, number 482, was posted on Thu Sep 11 '08 at 19:07:47
in reply to 5395f7047Tv-6828-1053-00.htm

Re: Problem in computing thermochemistry of small molecules (e.g. CO & CO2)

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you cannot use $FORCE DECOMP=.t. $END working in cartesian coordinates.
You must provide nonzero nzvar and specify z-matrix.

Regards,
Alex Granovsky


On Wed Sep 10 '08, Denis Zavelev wrote
--------------------------------------
>I'm trying to compute thermochemistry of some small molecules (namely CO, CO2 and O2) in order to estimate thermochemistry of the reaction studied.
>It doesn't matter what to give in input file: non-optimized geometry, optimized geometry without $HESS or with it - I am getting the following:
>=========
>*** ERROR *** ATTEMPT TO READ A DAF RECORD THAT WAS NEVER WRITTEN.
>RECORD NUMBER 49 OF LENGTH 0 DOES NOT EXIST.
>CHECK -PROG.DOC- FOR A LIST OF DIRECT ACCESS FILE CONTENTS
>=========
>
>
>Here's input file (for one of the attempts):
>=========
> $CONTRL RUNTYP=OPTIMIZE ECP=SBKJC COORD=CART DFTTYP=B3LYP NOSYM=1
> MAXIT=300 SCFTYP=RHF MULT=1 $END
> $SYSTEM TIMLIM=30000 MWORDS=40 $END
> $BASIS GBASIS=SBKJC $END
> $SCF DAMP=.t. SHIFT=.t. NCONV=6 $END
> $STATPT NSTEP=500 HSSEND=.t. OPTTOL=0.00001 $END
> $ELFLDG IEFLD=2 $END
> $ELDENS IEDEN=1 $END
> $FORCE TEMP=453 PRTSCN=.t. DECOMP=.t. $END
> $DATA
> CO
>C1
>CARBON 6.0 -1.225552804 -2.699077973 0.189065492
>OXYGEN 8.0 0.201717658 -2.610859894 0.193139549
> $END
>=========

>What am I doing wrong?


Message 50477d8426n-6830-629-00.htm, number 483, was posted on Fri Sep 12 '08 at 10:30:23
RESTART DATA FOR HESSIAN

Jamshid Najafpour
j.najafpour@gmail.com


I have made input file for restart data for hessian as attached but WINGAMESS package dont write it completely in output file.

This message contains the 5 kb attachment
[ N20-bowl1-C5v-3.inp ] Input file for restart data for hessian calculation


Message 4df668afbHW-6830-677+00.htm, number 484, was posted on Fri Sep 12 '08 at 11:17:04
in reply to 50477d8426n-6830-629-00.htm

Re: RESTART DATA FOR HESSIAN

Alex Granovsky
gran@classic.chem.msu.su


Hi,

note we do not discuss GAMESS(US)/WinGAMESS related issues on this forum.
This forum was specifically designated to discuss PC GAMESS related topics, which is a different program.
I'd suggest you to post your question to the GAMESS Google group.

Regards,
Alex Granovsky

On Fri Sep 12 '08, Jamshid Najafpour wrote
------------------------------------------
>I have made input file for restart data for hessian as attached but WINGAMESS package dont write it completely in output file.


Message 50477ad626n-6832-1064-00.htm, number 485, was posted on Sun Sep 14 '08 at 17:44:11
Computation with aug-cc-pVDZ basis set

Jamshid Najafpour
j.najafpour@gmail.com


Hi;

I want to use aug-cc-pVDZ basis set. How can I use acc-pvdz.lib file? and How can I put keyword in $BASIS?

Best Regards
Jamshid


Message dbf63a10So7-6832-1207-00.htm, number 486, was posted on Sun Sep 14 '08 at 20:07:12
What is more in this dramatic MISTAKE?

ganymede
wuxb07@lzu.cn


Dear Pro.
I have got a error in the following input;
! ?File created by MacMolPlt 7.2.1
$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=X3LYP MAXIT=200 MULT=1 $END
$SYSTEM TIMLIM=525600 MEMORY=1000000 $END
$SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
$STATPT OPTTOL=0.0001 NSTEP=200 $END
$DATA
Title
DNH 8

H ??1.0 ??0.00000 ??0.00000 ??0.97000
S ??
1 ?33.09250 ?00260941
2 ?5.022805 ?01966690
3 ?.9557020 ?09225256
4 ?.2027800 ?27352994
5 ?66434900 ?47654422
6 ?20824500 ?30155862
7 ?07288600 ?00826996
S ??
1 ?33.09250 ?.0039110
2 ?5.022805 ?.0303249
3 ?.9557020 ?.1971368
4 ?.2027800 ?.7675756
5 ?66434900 ?07121518
6 ?20824500 ?85518678
7 ?07288600 ?00011896
S ??
1 ?33.09250 ?00751951
2 ?5.022805 ?03413599
3 ?.9557020 ?73279322
4 ?.2027800 ?26520947
5 ?66434900 ?1.774089
6 ?20824500 ?.3616934
7 ?07288600 ?.0568172
S ??
1 ?33.09250 ?.0054170
2 ?5.022805 ?.1523083
3 ?.9557020 ?1.353979
4 ?.2027800 ?.5779532
5 ?66434900 ?2.337934
6 ?20824500 ?.0484932
7 ?07288600 ?01802187
S ??
1 ?33.09250 ?03125343
2 ?5.022805 ?1.426082
3 ?.9557020 ?.6956746
4 ?.2027800 ?1.077115
5 ?66434900 ?17350022
6 ?20824500 ?70730893
7 ?07288600 ?.7407261
S ??
1 ?33.09250 ?.0063455
2 ?5.022805 ?62069362
3 ?.9557020 ?1.053717
4 ?.2027800 ?.4475639
5 ?66434900 ?2.129268
6 ?20824500 ?.8952449
7 ?07288600 ?2.046450
P ?3
1 ?.0566000 ?31299579
2 ?76220000 ?74142300
3 ?19060000 ?08473191
P ?3
1 ?.0566000 ?1.079406
2 ?76220000 ?66739271
3 ?19060000 ?29174588
P ?3
1 ?.0566000 ?57446409
2 ?76220000 ?1.149113
3 ?19060000 ?.2264082

$END

In the inout file, the ANO basis for H is wrong!It is for He originally! I use it wrong for H, but, dramaticly, I run it in gamess and PC-GAMESS, it give a very accurate description of H2 ,giving a bond length 0.74206A, energy -1.1744391046 Ha!
I do not know why this happened? Is there something deep-seated reason in it ?


Message 4752619b8Nm-6836-58+00.htm, number 487, was posted on Thu Sep 18 '08 at 00:58:23
in reply to dbf63a10So7-6832-1207-00.htm

Re: What is more in this dramatic MISTAKE?

Jim Kress
jimkress_35@kressworks.com


You get a good result probably because you are using an uncontracted basis set and H/He basis sets do not differ significantly.

Jim

On Sun Sep 14 '08, ganymede wrote
---------------------------------
>Dear Pro.
> I have got a error in the following input;
>! ?File created by MacMolPlt 7.2.1
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=X3LYP MAXIT=200 MULT=1 $END
> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
> $STATPT OPTTOL=0.0001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>H ??1.0 ??0.00000 ??0.00000 ??0.97000
>S ??
>1 ?33.09250 ?00260941
>2 ?5.022805 ?01966690
>3 ?.9557020 ?09225256
>4 ?.2027800 ?27352994
>5 ?66434900 ?47654422
>6 ?20824500 ?30155862
>7 ?07288600 ?00826996
>S ??
>1 ?33.09250 ?.0039110
>2 ?5.022805 ?.0303249
>3 ?.9557020 ?.1971368
>4 ?.2027800 ?.7675756
>5 ?66434900 ?07121518
>6 ?20824500 ?85518678
>7 ?07288600 ?00011896
>S ??
>1 ?33.09250 ?00751951
>2 ?5.022805 ?03413599
>3 ?.9557020 ?73279322
>4 ?.2027800 ?26520947
>5 ?66434900 ?1.774089
>6 ?20824500 ?.3616934
>7 ?07288600 ?.0568172
>S ??
>1 ?33.09250 ?.0054170
>2 ?5.022805 ?.1523083
>3 ?.9557020 ?1.353979
>4 ?.2027800 ?.5779532
>5 ?66434900 ?2.337934
>6 ?20824500 ?.0484932
>7 ?07288600 ?01802187
>S ??
>1 ?33.09250 ?03125343
>2 ?5.022805 ?1.426082
>3 ?.9557020 ?.6956746
>4 ?.2027800 ?1.077115
>5 ?66434900 ?17350022
>6 ?20824500 ?70730893
>7 ?07288600 ?.7407261
>S ??
>1 ?33.09250 ?.0063455
>2 ?5.022805 ?62069362
>3 ?.9557020 ?1.053717
>4 ?.2027800 ?.4475639
>5 ?66434900 ?2.129268
>6 ?20824500 ?.8952449
>7 ?07288600 ?2.046450
>P ?3
>1 ?.0566000 ?31299579
>2 ?76220000 ?74142300
>3 ?19060000 ?08473191
>P ?3
>1 ?.0566000 ?1.079406
>2 ?76220000 ?66739271
>3 ?19060000 ?29174588
>P ?3
>1 ?.0566000 ?57446409
>2 ?76220000 ?1.149113
>3 ?19060000 ?.2264082

> $END

>In the inout file, the ANO basis for H is wrong!It is for He originally! I use it wrong for H, but, dramaticly, I run it in gamess and PC-GAMESS, it give a very accurate description of H2 ,giving a bond length 0.74206A, energy -1.1744391046 Ha!
>I do not know why this happened? Is there something deep-seated reason in it ?


Message 4752619b8Nm-6836-65+00.htm, number 488, was posted on Thu Sep 18 '08 at 01:05:19
in reply to 4245d5eaNOu-6820-6+00.htm

Re: Gaussian's O3LYP is wrong!

Jim Kress
jimkress_35@kressworks.com


I am shocked!, SHOCKED! to hear that there is a flaw in Gaussian :>}

Jim

Reference: the movie Casablanca

On Tue Sep 2 '08, Wenli Zou wrote
---------------------------------
>Thank you!

>You are right. I modified the source code of NWChem. Now the energy is -99.698055076796 a.u., being very close to the ones of PC-GAMESS and ORCA.

>Gaussian's O3LYP is indeed wrong. To correct it, the parameters of the correlation functionals must be specified explicitly by

># cc-pvtz IOp(3/74=-24) IOP(3/77=0813309262)

>or (for Gaussian 03.c.01 and higher)

># o3lyp/cc-pvtz IOP(3/77=0813309262)

>By this way, a correct energy of -99.6980592818 a.u. can be obtained.

>A test for F2 by O3LYP5/cc-pvtz with Grid=UltraFine.

>bond distance (Re) /Ang.
>1.3897 (G03's original O3LYP)
>1.3788 (G03's corrected O3LYP)
>1.3787 (PC-GAMESS)

>energy /a.u.
>-199.526380102 (G03's original O3LYP)
>-199.303011565 (G03's corrected O3LYP)
>-199.303013311 (PC-GAMESS)

>HOMO-LUMO gap /eV
>6.0 (G03's original O3LYP)
>6.1 (G03's corrected O3LYP)
>6.1 (PC-GAMESS)

>dissociation energy (De) /eV
>2.6 (G03's original O3LYP)
>2.5 (G03's corrected O3LYP)
>2.5 (PC-GAMESS)
>


Message 0ccb39185au-6838-295-00.htm, number 489, was posted on Sat Sep 20 '08 at 04:55:13
optimizing with PCGAMESS and frequencies with GAMESS_US

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


Three unrelated questions:
1. is it technically correct to optimize structure with PCGAMESS(because of the speed) and calculate frequencies with GAMESS-US (because of the cluster available). I would assume it's stil the same code, even though somewhat altered?
2. submitting batch of jobs to run in parallel on several PCs? Is it possible? How to organize it so next job doesn't overwrite previous output or punch file (I should ask G.Zhurko but thought might find an answer here)
3. which punch and nprint options should I choose to still get last $VEC punched out but not get 1Gb punch files or enormous output files in the end?

thanks

JOnas


Message 4df668afbHW-6838-736+00.htm, number 490, was posted on Sat Sep 20 '08 at 12:16:41
in reply to dbf63a10So7-6832-1207-00.htm

Re: What is more in this dramatic MISTAKE?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this problem is not related with PC GAMESS, but rather with MacMolPlt.

Please address your questions to its author.

Regards,
Alex Granovsky


On Sun Sep 14 '08, ganymede wrote
---------------------------------
>Dear Pro.
> I have got a error in the following input;
>! ?File created by MacMolPlt 7.2.1
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE DFTTYP=X3LYP MAXIT=200 MULT=1 $END
> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
> $STATPT OPTTOL=0.0001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>H ??1.0 ??0.00000 ??0.00000 ??0.97000
>S ??
>1 ?33.09250 ?00260941
>2 ?5.022805 ?01966690
>3 ?.9557020 ?09225256
>4 ?.2027800 ?27352994
>5 ?66434900 ?47654422
>6 ?20824500 ?30155862
>7 ?07288600 ?00826996
>S ??
>1 ?33.09250 ?.0039110
>2 ?5.022805 ?.0303249
>3 ?.9557020 ?.1971368
>4 ?.2027800 ?.7675756
>5 ?66434900 ?07121518
>6 ?20824500 ?85518678
>7 ?07288600 ?00011896
>S ??
>1 ?33.09250 ?00751951
>2 ?5.022805 ?03413599
>3 ?.9557020 ?73279322
>4 ?.2027800 ?26520947
>5 ?66434900 ?1.774089
>6 ?20824500 ?.3616934
>7 ?07288600 ?.0568172
>S ??
>1 ?33.09250 ?.0054170
>2 ?5.022805 ?.1523083
>3 ?.9557020 ?1.353979
>4 ?.2027800 ?.5779532
>5 ?66434900 ?2.337934
>6 ?20824500 ?.0484932
>7 ?07288600 ?01802187
>S ??
>1 ?33.09250 ?03125343
>2 ?5.022805 ?1.426082
>3 ?.9557020 ?.6956746
>4 ?.2027800 ?1.077115
>5 ?66434900 ?17350022
>6 ?20824500 ?70730893
>7 ?07288600 ?.7407261
>S ??
>1 ?33.09250 ?.0063455
>2 ?5.022805 ?62069362
>3 ?.9557020 ?1.053717
>4 ?.2027800 ?.4475639
>5 ?66434900 ?2.129268
>6 ?20824500 ?.8952449
>7 ?07288600 ?2.046450
>P ?3
>1 ?.0566000 ?31299579
>2 ?76220000 ?74142300
>3 ?19060000 ?08473191
>P ?3
>1 ?.0566000 ?1.079406
>2 ?76220000 ?66739271
>3 ?19060000 ?29174588
>P ?3
>1 ?.0566000 ?57446409
>2 ?76220000 ?1.149113
>3 ?19060000 ?.2264082

> $END

>In the inout file, the ANO basis for H is wrong!It is for He originally! I use it wrong for H, but, dramaticly, I run it in gamess and PC-GAMESS, it give a very accurate description of H2 ,giving a bond length 0.74206A, energy -1.1744391046 Ha!
>I do not know why this happened? Is there something deep-seated reason in it ?


Message 4df668afbHW-6838-738+00.htm, number 491, was posted on Sat Sep 20 '08 at 12:18:12
in reply to 50477ad626n-6832-1064-00.htm

Re: Computation with aug-cc-pVDZ basis set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

you can find basis set libraries at the "Downloads" section of the PC GAMESS site.

Regards,
Alex Granovsky

On Sun Sep 14 '08, Jamshid Najafpour wrote
------------------------------------------
>Hi;

>I want to use aug-cc-pVDZ basis set. How can I use acc-pvdz.lib file? and How can I put keyword in $BASIS?

>Best Regards
>Jamshid


Message 4df668afbHW-6840-1231+00.htm, number 492, was posted on Mon Sep 22 '08 at 20:31:12
in reply to 0ccb39185au-6838-295-00.htm

Re: optimizing with PCGAMESS and frequencies with GAMESS_US

Alexei Popov
alexei.a.popov@gmail.com


Dear Jonas,

1. This may or may not be correct, depending on your particular
problem. My view is that these two programs are now two quite a
different packages. I'd rather suggest you to ask cluster admins to
install parallel version of PC GAMESS on your cluster.

2. Yes it is. You may find useful some existing threads on this forum.

3. RTFM

-Alexei

On Sat Sep 20 '08, Jonas Baltrusaitis wrote
-------------------------------------------
>Three unrelated questions:
>1. is it technically correct to optimize structure with PCGAMESS(because of the speed) and calculate frequencies with GAMESS-US (because of the cluster available). I would assume it's stil the same code, even though somewhat altered?
>2. submitting batch of jobs to run in parallel on several PCs? Is it possible? How to organize it so next job doesn't overwrite previous output or punch file (I should ask G.Zhurko but thought might find an answer here)
>3. which punch and nprint options should I choose to still get last $VEC punched out but not get 1Gb punch files or enormous output files in the end?

>thanks

>JOnas


Message c6b48310SZP-6841-1078+00.htm, number 493, was posted on Tue Sep 23 '08 at 17:58:08
in reply to 0ccb39185au-6838-295-00.htm

Re: optimizing with PCGAMESS and frequencies with GAMESS_US

Slawomir Janicki
slawomir.janicki@comcast.net


Hi Jonas,

I use a very simple (and crude...) system of two batch files to run pcgamess on a WinXP cluster. You will need one of the implementations of MPI that is compatible with pcgamess. I use DeinoMPI.

I have a folder on all nodes c:\qc that stores all data and programs. PC-GAMESS is in c:\qc\pcgamess. The c:\qc folder is shared on all nodes to make the updates easier.

The main batch file (rg.bat) runs the queue. It is executed on the master node (I use a Remote Desktop Connection to do that). Each job is called with a line:

call c:\qc\gamess\runpgms inputfilename

To schedule the next job while the I simply append another line like that to rg.bat and save the file.

The runpgms.bat is the batch file that executes each job. This is the version for DeinoMPI:

"C:\Program Files\DeinoMPI\bin\mpiexec" -user janickis -hosts 6 CAM-W12329 2 CAM-W009879 CAM-I012342 2 CAM-I012267 2 CAM-I010915 2 CAM-I10003 -dir c:\qc c:\qc\gamess\pcgamess -i "c:\qc\%1.inp" -o "c:\qc\%1.log" -t c:\qc\t -r -f -p
move PUNCH "%1.punch"
move IRCDATA "%1.irc"
move t.0/FILE.47 "%1.47"
move t.0/FILE1.47 "%1.1.47"
move t.0/FILE2.47 "%1.2.47"
move t.0/FILE3.47 "%1.3.47"
move t.0/FILE4.47 "%1.4.47"
move t.0/FILE5.47 "%1.5.47"
del /q t.0\*.48

The batch calls mpiexec to run pcgamess on 10 CPUs (6 PCs). I tell pcgamess to use the inputfilename.inp file, put the results in inputfilename.log, use the default temporary folders c:\qc\t.n, and put all the text files in the c:\qc folder. After the job is finished I scavenge the punch and irc files. I also recover the inputfilename.47 files from NBO (note that even though the *.47 file is a text file in the current public version is is located in the temporary folder; Alex is working on changing this) and clean up the *.48 files from NBO manually (the other temporary files are overwritten by the next job).

After the job is done I have the inputfilename.inp, inputfilename.log, inputfilename.punch, inputfilename.irc, inputfilename.47, and inputfilename.1.47 files in the c:\qc folder on the master node to collect.

One of these days I will learn something more sophisticated...

Slawomir

On Sat Sep 20 '08, Jonas Baltrusaitis wrote
-------------------------------------------
>Three unrelated questions:
>1. is it technically correct to optimize structure with PCGAMESS(because of the speed) and calculate frequencies with GAMESS-US (because of the cluster available). I would assume it's stil the same code, even though somewhat altered?
>2. submitting batch of jobs to run in parallel on several PCs? Is it possible? How to organize it so next job doesn't overwrite previous output or punch file (I should ask G.Zhurko but thought might find an answer here)
>3. which punch and nprint options should I choose to still get last $VEC punched out but not get 1Gb punch files or enormous output files in the end?

>thanks

>JOnas


Message cfa13e4dJpb-6842-373-00.htm, number 494, was posted on Wed Sep 24 '08 at 06:19:05
sample input file

Alex
deckuofm@yahoo.com


Dear PC GAMESS users,
Could you please specify where I can find a sample input file for a simulation similar
to one in the attachment.

Thank you in advance

deckuofm@yahoo.com

This message contains the 2050 kb attachment
[ diels.mpeg ] MPEG of the Diels-Alder Reaction


Message 0ccb39185au-6845-199-00.htm, number 495, was posted on Sat Sep 27 '08 at 03:23:48
linear dependency with LANL2DZ

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


When I try to run calculations using LANL2DZ on transition metals and 6-31G** on others I get the linear dependency error. If, however, I run all electron calculations using 6-31G* on exactly the same input calculation runs very smoothly. What's the deal, why can;t I use pseudopotential? I got it from PNNL

I attached the input and output. Below is the error


*** FATAL ERROR ***
THE INPUT BASIS SET CONTAINS A LINEAR DEPENDENCE.
INPUT BASIS DIMENSION= 1016 SALC DIMENSION= 1016
THE SMALLEST EIGENVALUE OF THE OVERLAP MATRIX IS 1.670794E-48
THERE ARE 360 EIGENVALUES LESS THAN 1.00E-10
THE NUMBER OF LINEARLY INDEPENDENT ORBITALS KEPT IS 656

IF THERE ARE ANY SMALL EIGENVALUES:
CHECK YOUR OUTPUT FILE FOR DUPLICATE ATOMIC COORDINATES, OR
FOR DUPLICATE BASIS FUNCTIONS, OR FOR ZERO CONTRACTIONS
COEFFICIENTS, IF FOUND, THEN RESUBMIT THIS JOB. OTHERWISE,
IF THE LINEAR DEPENDENCE IS A RESULT OF USING MANY DIFFUSE
BASIS FUNCTIONS, YOU ARE OUT OF LUCK, THE SCF NEVER WILL
CONVERGE WITH THIS DEGREE OF LINEAR DEPENDENCE

IF THE SALC SPACE DOES NOT MATCH THE INPUT BASIS DIMENSION:
CHECK TO BE SURE YOU SCRUPULOUSLY FOLLOWED THE CONVENTION
FOR AXIS ORIENTATION IN $DATA (ESP. PERPENDICULAR C2 AXES)
AND THEN RESUBMIT THIS JOB.

WARNING! YOU ARE USING OUTDATED VERSION OF THE PC GAMESS!
PLEASE CHECK PC GAMESS HOMEPAGE FOR INFORMATION ON UPDATES!

This message contains the 45 kb attachment
[ input_1 ]


Message 8afd7ca0aKC-6846-64+00.htm, number 496, was posted on Sun Sep 28 '08 at 01:04:55
in reply to 0ccb39185au-6845-199-00.htm

Re: linear dependency with LANL2DZ

Richard
bonarlaw@liv.ac.uk


EMSL has recently had a problem L shells - look at your 6-31G(d) and compare it with the inbuilt basis...

Richard

On Sat Sep 27 '08, Jonas Baltrusaitis wrote
-------------------------------------------
>When I try to run calculations using LANL2DZ on transition metals and 6-31G** on others I get the linear dependency error. If, however, I run all electron calculations using 6-31G* on exactly the same input calculation runs very smoothly. What's the deal, why can;t I use pseudopotential? I got it from PNNL

>I attached the input and output. Below is the error
>
>
> *** FATAL ERROR ***
> THE INPUT BASIS SET CONTAINS A LINEAR DEPENDENCE.
> INPUT BASIS DIMENSION= 1016 SALC DIMENSION= 1016
> THE SMALLEST EIGENVALUE OF THE OVERLAP MATRIX IS 1.670794E-48
> THERE ARE 360 EIGENVALUES LESS THAN 1.00E-10
> THE NUMBER OF LINEARLY INDEPENDENT ORBITALS KEPT IS 656

> IF THERE ARE ANY SMALL EIGENVALUES:
> CHECK YOUR OUTPUT FILE FOR DUPLICATE ATOMIC COORDINATES, OR
> FOR DUPLICATE BASIS FUNCTIONS, OR FOR ZERO CONTRACTIONS
> COEFFICIENTS, IF FOUND, THEN RESUBMIT THIS JOB. OTHERWISE,
> IF THE LINEAR DEPENDENCE IS A RESULT OF USING MANY DIFFUSE
> BASIS FUNCTIONS, YOU ARE OUT OF LUCK, THE SCF NEVER WILL
> CONVERGE WITH THIS DEGREE OF LINEAR DEPENDENCE

> IF THE SALC SPACE DOES NOT MATCH THE INPUT BASIS DIMENSION:
> CHECK TO BE SURE YOU SCRUPULOUSLY FOLLOWED THE CONVENTION
> FOR AXIS ORIENTATION IN $DATA (ESP. PERPENDICULAR C2 AXES)
> AND THEN RESUBMIT THIS JOB.

> WARNING! YOU ARE USING OUTDATED VERSION OF THE PC GAMESS!
> PLEASE CHECK PC GAMESS HOMEPAGE FOR INFORMATION ON UPDATES!
>


Message 8afd7ca0aKC-6846-122-00.htm, number 497, was posted on Sun Sep 28 '08 at 02:02:33
simpler $BASIS

Richard
bonarlaw@liv.ac.uk


GAMESS-US has a uniquely clumsy way of assigning bases to atoms. How about a more generic type of $BASIS group as used in the GAMESS-UK and virtually every other program. Something along the lines of the following (make up your own...):

$NEWBAS
C N O
6-31+G(d)

H
6-31G(d)

H1
S 3
1 33.8650000 0.0254938
2 5.0947900 0.1903730
3 1.1587900 0.8521610
S 1
1 0.3258400 1.0000000
S 1
1 0.1027410 1.0000000
P 1
1 1.0000000 1.0000000
P 1
1 0.5000000 1.0000000

metal
favorite custom basis

etc..

$END

The coordinate block would be in $DATA, and element names would be remembered. I'm sure this is bottom of the list of priorities for the developers, but it would simplify inputs, particularly for beginners.

Richard


Message 4df668afbHW-6847-1333+00.htm, number 498, was posted on Mon Sep 29 '08 at 22:13:12
in reply to 8afd7ca0aKC-6846-122-00.htm

Re: simpler $BASIS

Alex Granovsky
gran@classic.chem.msu.su


Dear Richard,

The external basis set file feature, as well as PC GAMESS' input preprocessing could help to some degree.
However you are absolutely right that things must be done much more user-friendly.
I think we'll implement something similar to what you suggested in one of the next PC GAMESS releases.

-Alex



On Sun Sep 28 '08, Richard wrote
--------------------------------
>GAMESS-US has a uniquely clumsy way of assigning bases to atoms. How about a more generic type of $BASIS group as used in the GAMESS-UK and virtually every other program. Something along the lines of the following (make up your own...):

>$NEWBAS
>C N O
>6-31+G(d)

>H
>6-31G(d)

>H1
> S 3
> 1 33.8650000 0.0254938
> 2 5.0947900 0.1903730
> 3 1.1587900 0.8521610
> S 1
> 1 0.3258400 1.0000000
> S 1
> 1 0.1027410 1.0000000
> P 1
> 1 1.0000000 1.0000000
> P 1
> 1 0.5000000 1.0000000

>metal
>favorite custom basis

>etc..

>$END

>The coordinate block would be in $DATA, and element names would be remembered. I'm sure this is bottom of the list of priorities for the developers, but it would simplify inputs, particularly for beginners.

>Richard
>


Message 4df668afbHW-6847-1335-00.htm, number 499, was posted on Mon Sep 29 '08 at 22:15:37
Some news

Alex Granovsky
gran@classic.chem.msu.su


Dear forum users,

we have just published some interesting news in the "News" section.

Best regards,
Alex Granovsky


Message 4df668afbHW-6847-1335+00.htm, number 499, was edited on Mon Sep 29 '08 at 22:16:22
and replaces message 4df668afbHW-6847-1335-00.htm

Some news

Alex Granovsky
gran@classic.chem.msu.su


Dear forum users,

we have just published some interesting PC GAMESS-related news at the "News" section.

Best regards,
Alex Granovsky

[ This message was edited on Mon Sep 29 '08 by the author ]


Message 4df668afbHW-6847-1335+00.htm, number 499, was edited on Mon Sep 29 '08 at 22:16:48
Some news

Alex Granovsky
gran@classic.chem.msu.su


Dear forum users,

we have just published some interesting PC GAMESS-related news in the "News" section.

Best regards,
Alex Granovsky

[ This message was edited on Mon Sep 29 '08 by the author ]


Message c6b48310SZP-6849-84-00.htm, number 500, was posted on Wed Oct 1 '08 at 01:25:26
Basis set utility

Slawomir Janicki
slawomir.janicki@comcast.net


I have been running geometry optimization calculations on large systems (70+ atoms) with lots of polarization and diffuse functions. As expected, there were linear dependencies, and I had to manually remove some diffuse functions. At the restart of these jobs I had to manually copy the cartesian coordinates for each atom because PC-GAMESS doesn't punch out the $DATA with full basis set description yet.

I wrote a simple VB script to do this in MS Word and you are welcome to use it and, preferably, improve it.

To use it you need to place the script in normal.dot, create a new document, and paste the following:

- the cartesians from the end of the previous run
- the line "C1" with no additional characters in it
- the cartesians with the basis set description from the previous input file.

Each atom name needs to be unique, like C1, C2, etc.

The script will move the new coordinates into the lines in the old input and copy the result to the clipboard. All you need to do is to paste it into the new $DATA.

I attached a sample starting file and the result.

Enjoy,
Slawomir

This is the script:

Sub FormatFullBasisSet()
Dim myRange As Range
Selection.Collapse
Selection.StartOf Unit:=wdStory, Extend:=wdMove
While ActiveDocument.Paragraphs(1).Range.Words(1) <> "C1"
Set myRange = ActiveDocument.Paragraphs(1).Range
myWord = RTrim(myRange.Words(1))
If myWord = "" Then myWord = RTrim(myRange.Words(2))
MyPar = myRange.Text
myRange.Delete
Selection.Find.Execute FindText:=myWord, Forward:=True, MatchWholeWord:=True
Selection.StartOf Unit:=wdParagraph, Extend:=wdExtend
Selection.EndOf Unit:=wdParagraph, Extend:=wdExtend
If myWord = "C1" Then
Selection.Collapse Direction:=wdCollapseEnd
Selection.Find.Execute FindText:=myWord, Forward:=True, MatchWholeWord:=True
Selection.StartOf Unit:=wdParagraph, Extend:=wdExtend
Selection.EndOf Unit:=wdParagraph, Extend:=wdExtend
End If
Selection.Text = MyPar
Selection.Collapse
Selection.StartOf Unit:=wdStory, Extend:=wdMove
Wend
ActiveDocument.Select
ActiveDocument.Range.Copy
End Sub

This message contains the 28 kb attachment
[ input.doc ] input file


Message c6b48310SZP-6849-87+00.htm, number 501, was posted on Wed Oct 1 '08 at 01:27:03
in reply to c6b48310SZP-6849-84-00.htm

Re: Basis set utility

Slawomir Janicki
slawomir.janicki@comcast.net


And here is the output file.

Slawomir

On Wed Oct 1 '08, Slawomir Janicki wrote
----------------------------------------
>I have been running geometry optimization calculations on large systems (70+ atoms) with lots of polarization and diffuse functions. As expected, there were linear dependencies, and I had to manually remove some diffuse functions. At the restart of these jobs I had to manually copy the cartesian coordinates for each atom because PC-GAMESS doesn't punch out the $DATA with full basis set description yet.

>I wrote a simple VB script to do this in MS Word and you are welcome to use it and, preferably, improve it.

>To use it you need to place the script in normal.dot, create a new document, and paste the following:

>- the cartesians from the end of the previous run
>- the line "C1" with no additional characters in it
>- the cartesians with the basis set description from the previous input file.

>Each atom name needs to be unique, like C1, C2, etc.

>The script will move the new coordinates into the lines in the old input and copy the result to the clipboard. All you need to do is to paste it into the new $DATA.

>I attached a sample starting file and the result.

>Enjoy,
>Slawomir

>This is the script:

>Sub FormatFullBasisSet()
> Dim myRange As Range
> Selection.Collapse
> Selection.StartOf Unit:=wdStory, Extend:=wdMove
> While ActiveDocument.Paragraphs(1).Range.Words(1) <> "C1"
> Set myRange = ActiveDocument.Paragraphs(1).Range
> myWord = RTrim(myRange.Words(1))
> If myWord = "" Then myWord = RTrim(myRange.Words(2))
> MyPar = myRange.Text
> myRange.Delete
> Selection.Find.Execute FindText:=myWord, Forward:=True, MatchWholeWord:=True
> Selection.StartOf Unit:=wdParagraph, Extend:=wdExtend
> Selection.EndOf Unit:=wdParagraph, Extend:=wdExtend
> If myWord = "C1" Then
> Selection.Collapse Direction:=wdCollapseEnd
> Selection.Find.Execute FindText:=myWord, Forward:=True, MatchWholeWord:=True
> Selection.StartOf Unit:=wdParagraph, Extend:=wdExtend
> Selection.EndOf Unit:=wdParagraph, Extend:=wdExtend
> End If
> Selection.Text = MyPar
> Selection.Collapse
> Selection.StartOf Unit:=wdStory, Extend:=wdMove
> Wend
> ActiveDocument.Select
> ActiveDocument.Range.Copy
>End Sub

This message contains the 27 kb attachment
[ output.doc ] result


Message c8c31b02VlS-6855-980-00.htm, number 503, was posted on Tue Oct 7 '08 at 16:20:55
Basis set

Alexandre M Dias
alexandre.dias@unifenas.br


Hi all

How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?


Message c8c31b02VlS-6855-1076-00.htm, number 504, was posted on Tue Oct 7 '08 at 17:55:55
New user

Alexandre M Dias
alexandre.dias@unifenas.br


I am a new user of this forum and student of structure of diatomic molecules with PC Gamess. Patience with me !

I would like sharing with the members of this forum, a work what I published recently in the Magazine Physica B. He can be accessed by his digital code:
doi:10.1016/j.physb.2008.05.014


Message c8c31b02VlS-6855-1076+00.htm, number 504, was edited on Thu Oct 9 '08 at 02:28:31
and replaces message c8c31b02VlS-6855-1076-00.htm

New user

Alexandre M Dias
alexandre.dias@unifenas.br


I am a new user of this forum and student of structure of diatomic molecules with PC Gamess. Patience with me !
I would like sharing with the members of this forum, a work what I published recently in the Magazine Physica B. He can be accessed by his digital code: doi:10.1016/j.physb.2008.05.014 . The results was obtained with PC Gamess Project !

[ This message was edited on Thu Oct 9 '08 by the author ]


Message c8c31b02VlS-6855-980+00.htm, number 503, was edited on Thu Oct 9 '08 at 06:48:05
and replaces message c8c31b02VlS-6855-980-00.htm

Basis set

Alexandre M Dias
alexandre.dias@unifenas.br


Hi all

How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
I am asking because Gamess return a error message as incapacity to calculate integrals to M and N shells.

[ This message was edited on Thu Oct 9 '08 by the author ]


Message c8c31b02VlS-6855-980+00.htm, number 504, was edited on Thu Oct 9 '08 at 06:49:46
and replaces message c8c31b02VlS-6855-980-00.htm

Basis set

Alexandre M Dias
alexandre.dias@unifenas.br


Hi all

How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
I am asking because Gamess return a error message as incapacity to calculate integrals to M and N shells.

[ This message was edited on Thu Oct 9 '08 by the author ]


Message c8c31b02VlS-6855-980+00.htm, number 504, was edited on Thu Oct 9 '08 at 06:52:10
and replaces message c8c31b02VlS-6855-980-00.htm

Basis set

Alexandre M Dias
alexandre.dias@unifenas.br


Hi all

How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
I am asking because Gamess return a error message as incapacity to calculate integrals to M and N shells.

[ This message was edited on Thu Oct 9 '08 by the author ]


Message c8c31b02VlS-6855-980+00.htm, number 503, was edited on Thu Oct 9 '08 at 17:05:09
Basis set

Alexandre M Dias
alexandre.dias@unifenas.br


Hi all

How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
I am asking because the Gamess return a error message: incapacity to calculate integrals to M and N shells.

[ This message was edited on Thu Oct 9 '08 by the author ]


Message 4df668afbHW-6857-1340+00.htm, number 505, was posted on Thu Oct 9 '08 at 22:20:11
in reply to c8c31b02VlS-6855-980+00.htm

Re: Basis set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it seems you have syntax error in your input,
e.g. missed empty line after the end of input for each atom...

Regards,
Alex Granovsky


On Thu Oct 9 '08, Alexandre M Dias wrote
----------------------------------------
>Hi all

>How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
>I am asking because the Gamess return a error message: incapacity to calculate integrals to M and N shells.


Message c8c31b02VlS-6858-193+00.htm, number 506, was posted on Fri Oct 10 '08 at 03:13:06
in reply to 4df668afbHW-6857-1340+00.htm

Re^2: Basis set

Alexandre M Dias
alexandre.dias@unifenas.br



Excuse me, see my input file:

$CONTRL SCFTYP=RHF RUNTYP=energy UNITS=angs $END
$SYSTEM TIMLIM=15 MEMORY=2000000 $END
$GUESS GUESS=HUCKEL $END
$DATA
N2 .... Ground State cc-pVQZ
C1
N 7
N 7 0.0 0.0 0.549
S 9
1 45840.0000000 0.0000920
2 6868.0000000 0.0007170
3 1563.0000000 0.0037490
4 442.4000000 0.0155320
5 144.3000000 0.0531460
6 52.1800000 0.1467870
7 20.3400000 0.3046630
8 8.3810000 0.3976840
9 3.5290000 0.2176410
S 9
1 45840.0000000 -0.0000200
2 6868.0000000 -0.0001590
3 1563.0000000 -0.0008240
4 442.4000000 -0.0034780
5 144.3000000 -0.0119660
6 52.1800000 -0.0353880
7 20.3400000 -0.0800770
8 8.3810000 -0.1467220
9 3.5290000 -0.1163600
S 1
1 1.0540000 1.0000000
S 1
1 0.4118000 1.0000000
S 1
1 0.1552000 1.0000000
P 3
1 49.3300000 0.0055330
2 11.3700000 0.0379620
3 3.4350000 0.1490280
P 1
1 1.1820000 1.0000000
P 1
1 0.4173000 1.0000000
P 1
1 0.1428000 1.0000000
D 1
1 2.8370000 1.0000000
D 1
1 0.9680000 1.0000000
D 1
1 0.3350000 1.0000000
F 1
1 2.0270000 1.0000000
F 1
1 0.6850000 1.0000000
G 1
1 1.4270000 1.0000000

$END



On Thu Oct 9 '08, Alex Granovsky wrote
--------------------------------------
>Hi,

>it seems you have syntax error in your input,
>e.g. missed empty line after the end of input for each atom...

>Regards,
>Alex Granovsky
>
>
>On Thu Oct 9 '08, Alexandre M Dias wrote
>----------------------------------------
>>Hi all

>>How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
>>I am asking because the Gamess return a error message: incapacity to calculate integrals to M and N shells.


Message 4df668afbHW-6858-850+00.htm, number 507, was posted on Fri Oct 10 '08 at 14:09:44
in reply to 4245d5eaNOu-6820-6+00.htm

Re: Gaussian's O3LYP is wrong!

Alex Granovsky
gran@classic.chem.msu.su


Dear all,

I've copied below the message on this topic posted to CCL by
Gaussian's people. My opinion is that two possibilities mentioned
in CCL post simply define two different functionals. Certainly we
need clarification on this topic directly from O3LYP authors.
However, we've just modified the PC GAMESS sources to allow both
old-style (default) and Gaussian-style implementation. The PC GAMESS
v. 7.1.C (Firefly), to be released within a couple of days, has a
dedicated keyword allowing selection of the particular implementation
of O3LYP.

Best regards,
Alex Granovsky

----- Original Message -----
From: "Giovanni Scalmani giovanni-.-gaussian.com" owner-chemistry@ccl.net
To: "Granovsky, Alex, A. -id#4bg-" gran@classic.chem.msu.su
Sent: Friday, October 10, 2008 12:08 AM
Subject: CCL:G: Gaussian03 reproduces published O3LYP results.


>
> Sent to CCL by: Giovanni Scalmani [giovanni%gaussian.com]
>
> Our reference to reproduce OLYP and O3LYP results has been the following
> paper: W.-M. Hoe, A.J. Cohen, N.H. Handy, Chem. Phys. Letters 341 (2001) 319-328
> ("O3LYP paper").
>
> There are some ambiguities in the description of O3LYP in this paper. If we
> interpret these ambiguities one way, we get the same results as other programs
> but cannot reproduce the results in the paper. If we interpret them the other way,
> we can reproduce the results in the paper. Gaussian 03 revision D and later
> have been fixed to agree with the paper. The full details are given below.
>
>
> According to the paper, the OPTX exchange functional is defined as
>
> Rho^(4/3)*(A1*Cx + A2*Y^2) where Y = (g*x^2)/(1+g*x^2)
>
> and Rho is the density (alpha or beta), Cx is (3/4)*(6/PI)^(1/3), g=0.006 and
> x is the reduced density gradient x = |Grad_Rho|/(Rho^(4/3)). The two parameters
> are defined as follows A1=1.05151 and A2=1.43169. These values of the parameters
> allowed us to reproduce the OLYP results.
>
> Later in the paper, the authors give the values of the parameters for the
> O3LYP functional: A=0.1161 (% of exact exchange), B=0.9262 and C=0.8133. However,
> assuming that B is A1 and C is A2, produces results different from those in the
> paper. Using the values A1=0.9262 and A2=0.8133*1.43169 produces results in
> agreement with the paper.
>
> In Gaussian03 Rev. C.01 both the keywords OLYP and O3LYP were made available,
> but only OLYP gave results in agreement with the O3LYP paper, while O3LYP
> give wrong results. Setting IOp(3/77=0813309262) did not fixed the problem,
> but rather led to A1=0.9262 and A2=0.8133. Starting from Rev. D.01 the value of
> A2=0.8133*1.43169 was used by default and we were finally able to reproduce the
> published O3LYP results.
>
> Following are the details of some calculations that can be easily carried out
> with Gaussian03 to reproduce the results from Table 2 of the O3LYP paper. The
> agreement is very good for the H and He atoms, while is less than perfect for
> heavier atoms. However, given the larger absolute values of the energy and the
> uncertainty in the details of the calculations that were carried out to produce
> the published results (basis set, integration grid, SCF convergence, etc.), we
> are still confident that our implementation is correct.
>
> Nota that, G03 Rev. C.01 using IOp(3/77=0813309262) - which apparently agrees
> with other implementations of O3LYP - is clearly not able to reproduce the
> published data with the same accuracy.
>
> Basis set:
> - O3LYP paper : "PVQZ" (possibly from the CADPAC library, no literature
> reference is given)
> - following calculations: cc-pVQZ (5d,7f)
>
> Integration grid:
> - O3LYP paper : "high accuracy grid"
> - following calculations: int(grid=499974) (499 radial and 974 angular points)
>
> SCF convergence:
> - O3LYP paper : unspecified
> - following calculations: scf(tight) (RMS DP<1.0d-08, Max DP< 1.0d-06, DE<1.0d-6)
>
>
> Table 2: "Errors is total atomic energies (in hartrees) (theory-experiment)"
>
> Exact energies: H = -0.5000, He = -2.9037, Li = -7.4781, Be = -14.6674,
> F = -99.7339, Ne = -128.9376.
>
> H atom
> ======
> G03 Rev. G03 Table 2
> E.01 BLYP +0.0022 +0.0022
> E.01 OLYP +0.0016 +0.0015
> E.01 B3LYP (with VWN5) +0.0010 +0.0010
> E.01 O3LYP +0.0002 +0.0001
> C.01 O3LYP+IOp(3/77=0813309262) +0.0072
> E.01 B97-1 -0.0028 -0.0028
>
> He atom
> =======
> G03 Rev. G03 Table 2
> E.01 BLYP -0.0031 -0.0031
> E.01 OLYP -0.0035 -0.0036
> E.01 B3LYP (with VWN5) -0.0042 -0.0042
> E.01 O3LYP -0.0064 -0.0065
> C.01 O3LYP+IOp(3/77=0813309262) +0.0173
> E.01 B97-1 -0.0031 -0.0031
>
> Li atom
> =======
> G03 Rev. G03 Table 2
> E.01 BLYP -0.0033 -0.0045
> E.01 OLYP -0.0073 -0.0083
> E.01 B3LYP (with VWN5) -0.0038 -0.0045
> E.01 O3LYP -0.0095 -0.0102
> C.01 O3LYP+IOp(3/77=0813309262) +0.0302
> E.01 B97-1 -0.0070 -0.0078
>
> Be atom
> =======
> G03 Rev. G03 Table 2
> E.01 BLYP +0.0070 +0.0059
> E.01 OLYP +0.0008 -0.0004
> E.01 B3LYP (with VWN5) +0.0087 +0.0080
> E.01 O3LYP -0.0001 -0.0010
> C.01 O3LYP+IOp(3/77=0813309262) +0.0569
> E.01 B97-1 +0.0027 +0.0021
>
> F atom
> =======
> G03 Rev. G03 Table 2
> E.01 BLYP -0.0296 -0.0307
> E.01 OLYP -0.0108 -0.0123
> E.01 B3LYP (with VWN5) -0.0048 -0.0055
> E.01 O3LYP -0.0058 -0.0070
> C.01 O3LYP+IOp(3/77=0813309262) +0.1138
> E.01 B97-1 -0.0091 -0.0097
>
> Ne atom
> =======
> G03 Rev. G03 Table 2
> E.01 BLYP -0.0290 -0.0303
> E.01 OLYP -0.0120 -0.0138
> E.01 B3LYP (with VWN5) +0.0002 -0.0006
> E.01 O3LYP -0.0053 -0.0068
> C.01 O3LYP+IOp(3/77=0813309262) +0.1221
> E.01 B97-1 -0.0050 -0.0057
>
>
>
> -= This is automatically added to each message by the mailing script =-
> To recover the email address of the author of the message, please change
> the strange characters on the top line to the @ sign. You can also
> look up the X-Original-From: line in the mail header.
>
> E-mail to subscribers: CHEMISTRY@ccl.net or use:
> http://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/cgi-bin/ccl/send_ccl_message
>
> E-mail to administrators: CHEMISTRY-REQUEST@ccl.net or use
> http://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/cgi-bin/ccl/send_ccl_message
>
> Subscribe/Unsubscribe:
> http://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/chemistry/sub_unsub.shtml
>
> Before posting, check wait time at: http://www.ccl.nethttp://www.ccl.net
>
> Job: http://www.ccl.net/jobshttp://www.ccl.net/jobs
> Conferences: http://server.ccl.net/chemistry/announcements/conferences/http://server.ccl.net/chemistry/announcements/conferences/
>
> Search Messages: http://www.ccl.net/htdighttp://www.ccl.net/htdig (login: ccl, Password: search)
>
> If your mail bounces from CCL with 5.7.1 error, check:
> http://www.ccl.net/spammers.txthttp://www.ccl.net/spammers.txt
>
> RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/http://www.ccl.net/chemistry/aboutccl/instructions/
>
>
>



On Tue Sep 2 '08, Wenli Zou wrote
---------------------------------
>Thank you!

>You are right. I modified the source code of NWChem. Now the energy is -99.698055076796 a.u., being very close to the ones of PC-GAMESS and ORCA.

>Gaussian's O3LYP is indeed wrong. To correct it, the parameters of the correlation functionals must be specified explicitly by

># cc-pvtz IOp(3/74=-24) IOP(3/77=0813309262)

>or (for Gaussian 03.c.01 and higher)

># o3lyp/cc-pvtz IOP(3/77=0813309262)

>By this way, a correct energy of -99.6980592818 a.u. can be obtained.

>A test for F2 by O3LYP5/cc-pvtz with Grid=UltraFine.

>bond distance (Re) /Ang.
>1.3897 (G03's original O3LYP)
>1.3788 (G03's corrected O3LYP)
>1.3787 (PC-GAMESS)

>energy /a.u.
>-199.526380102 (G03's original O3LYP)
>-199.303011565 (G03's corrected O3LYP)
>-199.303013311 (PC-GAMESS)

>HOMO-LUMO gap /eV
>6.0 (G03's original O3LYP)
>6.1 (G03's corrected O3LYP)
>6.1 (PC-GAMESS)

>dissociation energy (De) /eV
>2.6 (G03's original O3LYP)
>2.5 (G03's corrected O3LYP)
>2.5 (PC-GAMESS)
>


Message 4df668afbHW-6858-854+00.htm, number 508, was posted on Fri Oct 10 '08 at 14:13:43
in reply to c8c31b02VlS-6858-193+00.htm

Re^3: Basis set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the correct input file is as follows:

 $CONTRL SCFTYP=RHF RUNTYP=energy UNITS=angs d5=.t. $END
 $SYSTEM TIMLIM=15 MEMORY=2000000 $END
 $GUESS GUESS=HUCKEL $END
 $DATA
 N2 .... Ground State cc-pVQZ
 C1
 N  7 0 0 -0.549
S  9
 1 45840.0000000       0.0000920    
 2  6868.0000000       0.0007170    
 3  1563.0000000       0.0037490    
 4  442.4000000       0.0155320    
 5  144.3000000       0.0531460    
 6   52.1800000       0.1467870    
 7   20.3400000       0.3046630    
 8   8.3810000       0.3976840    
 9   3.5290000       0.2176410    
S  9
 1 45840.0000000       -0.0000200    
 2  6868.0000000       -0.0001590    
 3  1563.0000000       -0.0008240    
 4  442.4000000       -0.0034780    
 5  144.3000000       -0.0119660    
 6   52.1800000       -0.0353880    
 7   20.3400000       -0.0800770    
 8   8.3810000       -0.1467220    
 9   3.5290000       -0.1163600    
S  1
 1   1.0540000       1.0000000    
S  1
 1   0.4118000       1.0000000    
S  1
 1   0.1552000       1.0000000    
P  3
 1   49.3300000       0.0055330    
 2   11.3700000       0.0379620    
 3   3.4350000       0.1490280    
P  1
 1   1.1820000       1.0000000    
P  1
 1   0.4173000       1.0000000    
P  1
 1   0.1428000       1.0000000    
D  1
 1   2.8370000       1.0000000    
D  1
 1   0.9680000       1.0000000    
D  1
 1   0.3350000       1.0000000    
F  1
 1   2.0270000       1.0000000    
F  1
 1   0.6850000       1.0000000    
G  1
 1   1.4270000       1.0000000

 N  7 0 0 0.549
S  9
 1 45840.0000000       0.0000920    
 2  6868.0000000       0.0007170    
 3  1563.0000000       0.0037490    
 4  442.4000000       0.0155320    
 5  144.3000000       0.0531460    
 6   52.1800000       0.1467870    
 7   20.3400000       0.3046630    
 8   8.3810000       0.3976840    
 9   3.5290000       0.2176410    
S  9
 1 45840.0000000       -0.0000200    
 2  6868.0000000       -0.0001590    
 3  1563.0000000       -0.0008240    
 4  442.4000000       -0.0034780    
 5  144.3000000       -0.0119660    
 6   52.1800000       -0.0353880    
 7   20.3400000       -0.0800770    
 8   8.3810000       -0.1467220    
 9   3.5290000       -0.1163600    
S  1
 1   1.0540000       1.0000000    
S  1
 1   0.4118000       1.0000000    
S  1
 1   0.1552000       1.0000000    
P  3
 1   49.3300000       0.0055330    
 2   11.3700000       0.0379620    
 3   3.4350000       0.1490280    
P  1
 1   1.1820000       1.0000000    
P  1
 1   0.4173000       1.0000000    
P  1
 1   0.1428000       1.0000000    
D  1
 1   2.8370000       1.0000000    
D  1
 1   0.9680000       1.0000000    
D  1
 1   0.3350000       1.0000000    
F  1
 1   2.0270000       1.0000000    
F  1
 1   0.6850000       1.0000000    
G  1
 1   1.4270000       1.0000000

$END

Regards,
Alex Granovsky


On Fri Oct 10 '08, Alexandre M Dias wrote
-----------------------------------------
>>Excuse me, see my input file:

>$CONTRL SCFTYP=RHF RUNTYP=energy UNITS=angs $END
> $SYSTEM TIMLIM=15 MEMORY=2000000 $END
> $GUESS GUESS=HUCKEL $END
> $DATA
> N2 .... Ground State cc-pVQZ
> C1
> N 7
> N 7 0.0 0.0 0.549
>S 9
> 1 45840.0000000 0.0000920
> 2 6868.0000000 0.0007170
> 3 1563.0000000 0.0037490
> 4 442.4000000 0.0155320
> 5 144.3000000 0.0531460
> 6 52.1800000 0.1467870
> 7 20.3400000 0.3046630
> 8 8.3810000 0.3976840
> 9 3.5290000 0.2176410
>S 9
> 1 45840.0000000 -0.0000200
> 2 6868.0000000 -0.0001590
> 3 1563.0000000 -0.0008240
> 4 442.4000000 -0.0034780
> 5 144.3000000 -0.0119660
> 6 52.1800000 -0.0353880
> 7 20.3400000 -0.0800770
> 8 8.3810000 -0.1467220
> 9 3.5290000 -0.1163600
>S 1
> 1 1.0540000 1.0000000
>S 1
> 1 0.4118000 1.0000000
>S 1
> 1 0.1552000 1.0000000
>P 3
> 1 49.3300000 0.0055330
> 2 11.3700000 0.0379620
> 3 3.4350000 0.1490280
>P 1
> 1 1.1820000 1.0000000
>P 1
> 1 0.4173000 1.0000000
>P 1
> 1 0.1428000 1.0000000
>D 1
> 1 2.8370000 1.0000000
>D 1
> 1 0.9680000 1.0000000
>D 1
> 1 0.3350000 1.0000000
>F 1
> 1 2.0270000 1.0000000
>F 1
> 1 0.6850000 1.0000000
>G 1
> 1 1.4270000 1.0000000

>$END
>
>
>
>On Thu Oct 9 '08, Alex Granovsky wrote
>--------------------------------------
>>Hi,

>>it seems you have syntax error in your input,
>>e.g. missed empty line after the end of input for each atom...

>>Regards,
>>Alex Granovsky
>>
>>
>>On Thu Oct 9 '08, Alexandre M Dias wrote
>>----------------------------------------
>>>Hi all

>>>How I do N2 single point RHF calculation without symmetry (C1) with cc-pVQZ basis set ?
>>>I am asking because the Gamess return a error message: incapacity to calculate integrals to M and N shells.


Message 829e844cbK2-6858-866-00.htm, number 509, was posted on Fri Oct 10 '08 at 14:25:55
Selection of a Basis set for heavy metals like Ag and Au

Mahesh Bhatt
mbhatt54@yahoo.com


Dear Sir,
When I used the basis set 6-31G for Ag(111) and Au(111) surfaces having 13 atoms, there was a warning in pcgamess calculation that an illegal extended basis set was requested. Thats why calculation was terminated. This basis set was OK for lighter metals like Ca, Mg and Al. Can I use a basis set MINI for Ag and Au surfaces? If yes , its OK. But if no, then what is other option?
Sincerely
Mahesh Bhatt
University of Tsukuba
Japan

Message 4df668afbHW-6858-894+00.htm, number 510, was posted on Fri Oct 10 '08 at 14:56:17
in reply to 829e844cbK2-6855-718-00.htm

Re: SCF unconvergence

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I've attached sample input and output files for your system.
The idea is to converge SCF using MINI basis set and then use
SCF vectors as a starting guess for extended basis set calculations.

Hope this helps.

Regards,
Alex Granovsky


On Tue Oct 7 '08, Mahesh Bhatt wrote
------------------------------------
>Dear Sir,
>I would like to request you some querries about pcgamess.If SCF can not converge after many iterations for any particular structure of an interface between organic semiconductor and metal, then can we have any options in input to control its convergence? Are there all energy levels as well as electron density with each iteration can be read from output file given in pcgamess? By usual calculation, we could not get convergence for larger distance for any particular interface. For example, we could not get convergence for Bathocuproine (BCP) - Calcium (Ca) 111 surface for distances greater than 4.20 angstron.How can we find the origin of this unconvergence in pcgamess? From output file or others?Plz sir give me reply as well as possible.
>Sincerely
>Mahesh Bhatt

This message contains the 2456 kb attachment
[ bcp.rar ] BCP results


Message 4df668afbHW-6858-906+00.htm, number 511, was posted on Fri Oct 10 '08 at 15:06:08
in reply to cfa13e4dJpb-6842-373-00.htm

Re: sample input file

Alex Granovsky
gran@classic.chem.msu.su


Hi,

first, for given reactants and products, you need to locate
transition state (TS). See the PC GAMESS manual on how to do this.

Next, you should calculare IRC starting from TS and connecting
reactants and products. Once again, see the manual for more information!

After you calculate IRC, perform calculations on a set of
representative points on IRC to get information on molecular
orbitals and/or overall electronic density.

Finally, visualize results using any visualization software
compatible with PC GAMESS. Save the plots of molecules and
orbitals at different geometries into the separate gif files
and convert them into animated gif or whatever you want using
graphical/video editors.

Regards,
Alex Granovsky


On Wed Sep 24 '08, Alex wrote
-----------------------------
>Dear PC GAMESS users,
>Could you please specify where I can find a sample input file for a simulation similar
>to one in the attachment.

>Thank you in advance

>deckuofm@yahoo.com
>


Message dcd1093cNCe-6859-815-00.htm, number 512, was posted on Sat Oct 11 '08 at 13:35:43
Solvation effect computation by EFP method

Fumihito Mohri
qzg00677@nifty.com


Dear All
Now, I am interested in the application of EFP method to solvent effect computation,
since I think that solvent effect computation by EFP is more sophisticated than PCM.
So, I read the description of $EFRAG in INPUT.pdf, and then I found the following
explanation.
POSITION=OPTIMIZE Allows full optimization within the
ab initio part, and optimization of
the rotational and translational
motions of each fragment. (default)

In EFP method, each "fragment molecule" is treated as multi-poles (a set of point
charges, one dipole or quadrupole etc). But, I consider with Earnshaw's theorem,
in order to obtain the optimization of the rotational and translational motions,
I should determine the form of exchange repulsions between fragment molecules and
between fragment molecules and ab-initio molecules. Then, I have a question: Should
I define the REPULSIVE POTENTIAL in $FRAGNAME and the inter-fragment repulsive potential in $FRGRPL?

Best regards,
Fumihito Mohri


Message dcdb5a17NCe-6860-264+00.htm, number 513, was posted on Sun Oct 12 '08 at 04:24:14
in reply to 829e844cbK2-6858-866-00.htm

Re: Selection of a Basis set for heavy metals like Ag and Au

Fumihito Mohri
qzg00677@nifty.com


Dear Mahesh Bhatt
First, please read the description of $BASIS in INPUT.pdf .
Second, you can get various basis functions in the following web site.
https://bse.pnl.gov/bse/portal

>University of Tsukuba
Oh, you are in Japan. Then, the following web site written in Japanese may be helpful
for you.
http://blog.3016.net/2006/03/post_2.html#more

Best regards,
Fumihito Mohri


Message d9da3625HWK-6861-687-00.htm, number 514, was posted on Mon Oct 13 '08 at 11:28:23
Basis set for la-lu

Siavash Riahi
riahisv@ipe.ut.ac.ir


Dear all
I perform various calculations using Lanthanum metals with organic molecules attached to them. which basis set in gamess was used? and How can I put keyword in input file?. The sample file was attached. Is correct?
thanks a lot
siavash

This message contains the 2 kb attachment
[ S1-La3+11.inp ] sample file


Message 4df668afbHW-6861-1192+00.htm, number 515, was posted on Mon Oct 13 '08 at 19:52:32
in reply to dcd1093cNCe-6859-815-00.htm

Re: Solvation effect computation by EFP method

Alex Granovsky
gran@classic.chem.msu.su


Dear Dr. Fumihito Mohri,

Indeed, if you omit all terms but pure electrostatics,
this most likely results in Coulomb collapse of your system.
In any case, you will not be able to find the optimal geometry.
Thus, you need both the ab-initio-EFP and EFP-EFP Pauli repulsion terms.

Unfortunately, the PC GAMESS does not allow fully automated
calculation of these terms for arbitrary fragment.

There exists internally stored standard fragment for H2O molecule,
which can be used in calculations and you do not need to specify
any additional parameters for it.

However, for other solvents, you can obtain all the required
electrostatic terms using runtyp=makefp, but not the parametrization
of repulsion terms. Thus, you need to adjust some (initially guessed)
parameters of repulsion manually using a set of QC calculations like
Morokuma decomposition, etc...

As far as I remember, there should be some program available
which was developed and used by the authors of the EFP model
to fit repulsive terms for water and other molecules.


Best regards,
Alex Granovsky


On Sat Oct 11 '08, Fumihito Mohri wrote
---------------------------------------
>Dear All
>Now, I am interested in the application of EFP method to solvent effect computation,
>since I think that solvent effect computation by EFP is more sophisticated than PCM.
> So, I read the description of $EFRAG in INPUT.pdf, and then I found the following
>explanation.
> POSITION=OPTIMIZE Allows full optimization within the
> ab initio part, and optimization of
> the rotational and translational
> motions of each fragment. (default)

>In EFP method, each "fragment molecule" is treated as multi-poles (a set of point
> charges, one dipole or quadrupole etc). But, I consider with Earnshaw's theorem,
>in order to obtain the optimization of the rotational and translational motions,
>I should determine the form of exchange repulsions between fragment molecules and
>between fragment molecules and ab-initio molecules. Then, I have a question: Should
>I define the REPULSIVE POTENTIAL in $FRAGNAME and the inter-fragment repulsive potential in $FRGRPL?

> Best regards,
> Fumihito Mohri
>


Message 542fef5f26n-6863-62+00.htm, number 516, was posted on Wed Oct 15 '08 at 01:02:35
in reply to d9da3625HWK-6861-687-00.htm

Re: Basis set for la-lu

Jamshid Najafpour
j.najafpour@gmail.com


Hi;

If you will put NGAUSS=6 in $BASIS line, your basis set is 6-311G.

Best Whishes
Jamshid
On Mon Oct 13 '08, Siavash Riahi wrote
--------------------------------------
>Dear all
>I perform various calculations using Lanthanum metals with organic molecules attached to them. which basis set in gamess was used? and How can I put keyword in input file?. The sample file was attached. Is correct?
>thanks a lot
>siavash


Message 829e844cbK2-6863-871-00.htm, number 517, was posted on Wed Oct 15 '08 at 14:30:55
Selection of basis set

Mahesh Bhatt
mbhatt54@yahoo.com


Dear all
I performed various calculations using different metals like Ca, Mg, Al, Ag and Au with organic molecules
attached to them. I have used N Gauss and a basis set MINI, because when I used 6-31G or 6-311G, the pcgamess gave the warning that illegal extended basis sets requested. As a result, the calculation was abnormally stopped. In such a case, which basis set in pcgamess can be used? Does B3LYP5 with MINI basis set give accurate calculation for organic molecule adsorbed on metal Ca, Mg, Al, Ag and Au 111 surfaces?
thanks a lot
Mahesh Bhatt
University of Tsukuba
Japan

Message d4c0f196bHW-6863-1010+00.htm, number 518, was posted on Wed Oct 15 '08 at 16:50:26
in reply to 829e844cbK2-6863-871-00.htm

Re: Selection of basis set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it is beter to use ECP basis sets for your metals.

Regards,
Alex Granovsky

On Wed Oct 15 '08, Mahesh Bhatt wrote
-------------------------------------
>Dear all
>I performed various calculations using different metals like Ca, Mg, Al, Ag and Au with organic molecules
>attached to them. I have used N Gauss and a basis set MINI, because when I used 6-31G or 6-311G, the pcgamess gave the warning that illegal extended basis sets requested. As a result, the calculation was abnormally stopped. In such a case, which basis set in pcgamess can be used? Does B3LYP5 with MINI basis set give accurate calculation for organic molecule adsorbed on metal Ca, Mg, Al, Ag and Au 111 surfaces?
>thanks a lot
>Mahesh Bhatt
>University of Tsukuba
>Japan


Message dcd1097bNCe-6863-1023+00.htm, number 519, was posted on Wed Oct 15 '08 at 17:03:00
in reply to 4df668afbHW-6861-1192+00.htm

Re^2: Solvation effect computation by EFP method

Fumihito Mohri
qzg00677@nifty.com


Dear Prof. Granovsky
Thank you for your answer.

>Thus, you need both the ab-initio-EFP and EFP-EFP Pauli repulsion terms.
Oh, this is "as I expected".

>Thus, you need to adjust some (initially guessed) parameters of
>repulsion manually using a set of QC calculations like Morokuma
>decomposition, etc...

I see.

>As far as I remember, there should be some program available
>which was developed and used by the authors of the EFP model
>to fit repulsive terms for water and other molecules.
I will seek the programs.
Does someone know the programs? -> All the users.

Best regards,
Fumihito Mohri


Message d4c0f196bHW-6863-1247+00.htm, number 518, was edited on Wed Oct 15 '08 at 20:47:30
and replaces message d4c0f196bHW-6863-1010+00.htm

Re: Selection of basis set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

it is better to use ECP basis sets to model your systems.

Regards,
Alex Granovsky

On Wed Oct 15 '08, Mahesh Bhatt wrote
-------------------------------------
>Dear all
>I performed various calculations using different metals like Ca, Mg, Al, Ag and Au with organic molecules
>attached to them. I have used N Gauss and a basis set MINI, because when I used 6-31G or 6-311G, the pcgamess gave the warning that illegal extended basis sets requested. As a result, the calculation was abnormally stopped. In such a case, which basis set in pcgamess can be used? Does B3LYP5 with MINI basis set give accurate calculation for organic molecule adsorbed on metal Ca, Mg, Al, Ag and Au 111 surfaces?
>thanks a lot
>Mahesh Bhatt
>University of Tsukuba
>Japan

[ This message was edited on Wed Oct 15 '08 by the author ]


Message 5b7ce244PMq-6864-40+00.htm, number 520, was posted on Thu Oct 16 '08 at 00:40:21
in reply to d9da3625HWK-6861-687-00.htm

Re: Basis set for la-lu

Vitaliy Gorbenko
vgorbenko@ukr.net


You have some mistakes. If you are choosing RHF then don't apply MULT=2. Next, you can't use the basis set 6-311G for La.
For your DATA you may use pseudopotential SBKJC.
With best regards,

Dr. Vitaliy Gorbenko


On Mon Oct 13 '08, Siavash Riahi wrote
--------------------------------------
>Dear all
>I perform various calculations using Lanthanum metals with organic molecules attached to them. which basis set in gamess was used? and How can I put keyword in input file?. The sample file was attached. Is correct?
>thanks a lot
>siavash


Message 8afd7ca0aKC-6864-72-00.htm, number 521, was posted on Thu Oct 16 '08 at 01:11:51
feature request: DFT-D

Richard
bonarlaw@liv.ac.uk


Would the developers consider implementing Grimme's empirical correction? It would make good use of pc-gamess's efficiency.

Thanks

Richard


Message 4df668afbHW-6865-1277-00.htm, number 522, was posted on Fri Oct 17 '08 at 21:17:22
Release of the PC GAMESS v. 7.1.C (Firefly)

Alex Granovsky
gran@classic.chem.msu.su


Dear PC GAMESS users,

The new official PC GAMESS release called "Firefly", version 7.1.C
i.e., 7.1.12 decimal), build # 5014, has just been made publicly
available at the Downloads section of the PC GAMESS Homepage.

The downloads of older PC GAMESS releases from ISU are no longer
supported. Note you need to get the new installation password
finally! only the single password!) from PC GAMESS/Firefly Project
Team for these binaries. The complete documentation on the new
features will be published at the PC GAMESS/Firefly web server
within a couple of weeks or so.

We hope you'll enjoy the new version of our software.

Best regards,
Alex Granovsky


Message 5395f7047Tv-6870-1177-00.htm, number 523, was posted on Wed Oct 22 '08 at 19:36:46
B97-1 functional

Denis Zavelev
denis.zavelev@gmail.com


Hello!

I'd like to try B97-1 DFT functional, but it seems that it's not available in Firefly. Is there any way to use it with Firefly?


Message ca3636f00sq-6870-1290-00.htm, number 524, was posted on Wed Oct 22 '08 at 21:30:33
anharmonic frequency calculations

Prasenjit Pandey
ppchem@gmail.com


Can anharmonic frequency calculations be done by PC GAMESS? If yes, then how?

Message 4df668afbHW-6873-1301+00.htm, number 525, was posted on Sat Oct 25 '08 at 21:40:35
in reply to 8afd7ca0aKC-6864-72-00.htm

Re: feature request: DFT-D

Alex Granovsky
gran@classic.chem.msu.su


Dear Richard,

we are currently considering this possibility.

Regards,
Alex

On Thu Oct 16 '08, Richard wrote
--------------------------------
>Would the developers consider implementing Grimme's empirical correction? It would make good use of pc-gamess's efficiency.

>Thanks

>Richard


Message 4df668afbHW-6873-1303+00.htm, number 526, was posted on Sat Oct 25 '08 at 21:42:57
in reply to 5395f7047Tv-6870-1177-00.htm

Re: B97-1 functional

Alex Granovsky
gran@classic.chem.msu.su


Hi,

at present, meta-GGA type functionals are not supported.
See previous discussions on this topic on forum for more information.

Regards,
Alex Granovsky




On Wed Oct 22 '08, Denis Zavelev wrote
--------------------------------------
>Hello!

>I'd like to try B97-1 DFT functional, but it seems that it's not available in Firefly. Is there any way to use it with Firefly?


Message 4df668afbHW-6873-1315+00.htm, number 527, was posted on Sat Oct 25 '08 at 21:54:37
in reply to ca3636f00sq-6870-1290-00.htm

Re: anharmonic frequency calculations

Alex Granovsky
gran@classic.chem.msu.su


Hi,

On Wed Oct 22 '08, Prasenjit Pandey wrote
-----------------------------------------
>Can anharmonic frequency calculations be done by PC GAMESS? If yes, then how?

This cannot be done with PC GAMESS/Firefly.
The reason is very simple - we prefer not to mix
electronic and (ro)vibrational problems in the
framework of single package. The (ro)vibrational
problem is the quite a separate and usually very
complicated task, at least if is being solved
without making serious, sometimes very crude
approximations. It often requires an individual
approach for each molecular system of interest.

The PC GAMESS/Firefly is primary the Electronic Structure package.
It does can be used as the computational server to
compute important parts of PES for subsequent use.
However, the (ro)vibrational problem itself should
be solved by a separate program.

Regards,
Alex Granovsky





Message 829e844cbK2-6874-608-00.htm, number 528, was posted on Sun Oct 26 '08 at 10:08:14
Basis set for Au111 surface

Mahesh Bhatt
mbhatt54@yahoo.com


Dear all,
I have construced Au 111 surface by adding 13 Au atoms and then tried to calculate energy using B3LYP5 DFT parameter with basis sets STO-2G,3G,4G,5G,6G and 3-21G, 4-31G, 6-31G, 6-311G. But in PCGAMESS, there was a warning that illegal extended basis sets not requested and calculation terminated abnormally. In such a case, which basis set is available for Au surface in PCGAMESS? I have then used MINI basis set for calculation. But in case of a MINI basis set, there is zero charge transfer in Carbon and Nitrogen atoms when interacting an organic molecule with Au surface. This is a strange, isnt it?
Is there any solution in PCGAMESS?
Second thing, I cant install PC GAMESS/Firefly software in my PC like PCGAMESS. In PCGAMESS, I got pcg701 which could be easily installed after entering the given password. But in PC GAMESS/Firefly , I got pcg71c RAR file, which can not be installed because it can not be extracted in pc windows. I would like to install it, because my project incudes electronic structure calculation only. So I hope may be this software useful for me.
Thanking you all
Mahesh Bhatt
University of Tsukuba
Japan

Message 589dd852bQn-6875-733+00.htm, number 529, was posted on Mon Oct 27 '08 at 12:12:34
in reply to 829e844cbK2-6874-608-00.htm

Re: Basis set for Au111 surface

Pedro Silva
pedros@ufp.pt


On Sun Oct 26 '08, Mahesh Bhatt wrote
-------------------------------------
Second thing, I cant install PC GAMESS/Firefly software in my PC like PCGAMESS. In PCGAMESS, I got pcg701 which could be easily installed after entering the given password. But in PC GAMESS/Firefly , I got pcg71c RAR file, which can not be installed because it can not be extracted in pc windows. I would like to install it, because my project incudes electronic structure calculation only. So I hope may be this software useful for me.
---------------------------------

RAR is a compression format, like .z, .gz. .zip or .arj
You may download the WinRar utility from
www.rarlab.com


Message 829e844cbK2-6876-385-00.htm, number 530, was posted on Tue Oct 28 '08 at 06:24:39
Selection of a basis set for Au111 surface

Mahesh Bhatt
mbhatt54@yahoo.com


Dear all,
I have construced Au 111 surface by adding 13 Au atoms and then tried to calculate energy using B3LYP5 DFT parameter with basis sets STO-2G,3G,4G,5G,6G and 3-21G, 4-31G, 6-31G, 6-311G. But in PCGAMESS, there was a warning that illegal extended basis sets not requested and calculation terminated abnormally. In such a case, which basis set is available for Au surface in PCGAMESS? I have then used MINI basis set for calculation. But in case of a MINI basis set, there is zero charge transfer in Carbon and Nitrogen atoms when interacting an organic molecule with Au surface. This is a strange, isnt it?
Is there any solution in PCGAMESS?
Thanking you all
Mahesh Bhatt
University of Tsukuba
Japan

Message 83d707ec579-6876-817-00.htm, number 531, was posted on Tue Oct 28 '08 at 13:36:56
Fatal error no. 0x00020070 in sub PACKWRIT on unit 8

Sibo
siboster@gmail.com


Hi, I am new in Firefly--all I've successfully run is an optimization of 1-methylcyclohexene. Now I'm being a bit more ambitious and would like to optimize the structure of a poly-aromatic ligand with some copper ions. However, after determining the symmetries for the initial guess orbitals, my calculation crapped out. Can anyone tell me what this error means and how to fix it?

Here's my input:
$contrl scftyp=rhf dfttyp=b3lyp5 runtyp=optimize nprint=-5 $end
$system mwords=64 timlim=600 $end
$basis gbasis=TZV ndfunc=1 npfunc=1 $end
$statpt nstep=50 $end
[data, atomic coordinates]

And here's the end of my output (� corresponds to whitespace--I'm not sure why when I copy and paste I get this odd symbol...):
SYMMETRIES FOR INITIAL GUESS ORBITALS FOLLOW. � BOTH SET(S).
� 203 ORBITALS ARE OCCUPIED ( � 75 CORE ORBITALS).
� �76=A � � � 77=A � � � 78=A � � � 79=A � � � 80=A � � � 81=A � � � 82=A �
� �83=A � � � 84=A � � � 85=A � � � 86=A � � � 87=A � � � 88=A � � � 89=A �
� �90=A � � � 91=A � � � 92=A � � � 93=A � � � 94=A � � � 95=A � � � 96=A �
� �97=A � � � 98=A � � � 99=A � � �100=A � � �101=A � � �102=A � � �103=A �
� 104=A � � �105=A � � �106=A � � �107=A � � �108=A � � �109=A � � �110=A �
� 111=A � � �112=A � � �113=A � � �114=A � � �115=A � � �116=A � � �117=A �
� 118=A � � �119=A � � �120=A � � �121=A � � �122=A � � �123=A � � �124=A �
� 125=A � � �126=A � � �127=A � � �128=A � � �129=A � � �130=A � � �131=A �
� 132=A � � �133=A � � �134=A � � �135=A � � �136=A � � �137=A � � �138=A �
� 139=A � � �140=A � � �141=A � � �142=A � � �143=A � � �144=A � � �145=A �
� 146=A � � �147=A � � �148=A � � �149=A � � �150=A � � �151=A � � �152=A �
� 153=A � � �154=A � � �155=A � � �156=A � � �157=A � � �158=A � � �159=A �
� 160=A � � �161=A � � �162=A � � �163=A � � �164=A � � �165=A � � �166=A �
� 167=A � � �168=A � � �169=A � � �170=A � � �171=A � � �172=A � � �173=A �
� 174=A � � �175=A � � �176=A � � �177=A � � �178=A � � �179=A � � �180=A �
� 181=A � � �182=A � � �183=A � � �184=A � � �185=A � � �186=A � � �187=A �
� 188=A � � �189=A � � �190=A � � �191=A � � �192=A � � �193=A � � �194=A �
� 195=A � � �196=A � � �197=A � � �198=A � � �199=A � � �200=A � � �201=A �
� 202=A � � �203=A � � �204=A � � �205=A � � �206=A � � �207=A � � �208=A �
� 209=A � � �210=A � � �211=A � � �212=A � � �213=A �
...... END OF INITIAL ORBITAL SELECTION ......

CPU � � � �TIME: � STEP = � � 20.95 , �TOTAL = � � � 28.1 SECONDS ( � �0.5 MIN)
WALL CLOCK TIME: � STEP = � � 30.19 , �TOTAL = � � � 38.8 SECONDS ( � �0.6 MIN)
CPU UTILIZATION: � STEP = � � 69.40%, �TOTAL = � � �72.42%
FSF: fatal error no. 0x00020070 in sub PACKWRIT on unit � 8

CPU � � � �TIME: � STEP = � 5799.57 , �TOTAL = � � 5827.7 SECONDS ( � 97.1 MIN)
WALL CLOCK TIME: � STEP = � 7388.83 , �TOTAL = � � 7427.6 SECONDS ( �123.8 MIN)
CPU UTILIZATION: � STEP = � � 78.49%, �TOTAL = � � �78.46%

� �13378439 WORDS OF � �DYNAMIC MEMORY USED
EXECUTION OF PC GAMESS TERMINATED ABNORMALLY AT �1:14:51 LT �28-OCT-2008


Message 808be225Vl5-6876-961-00.htm, number 532, was posted on Tue Oct 28 '08 at 16:01:02
about $EFIELD group

xuefei
xuxuefei@gmail.com


Dear all,

Now, I am tring to use the $EFIELD group to study the influence of an external electric field on the molecule. I wonder what is the limit of the electric field strength? What theoretical model is used in calculations? Could you please intoduce some reference about that to me?

Thanks a lot!

Best wishes,
xuefei


Message d4c0f196bHW-6877-941+00.htm, number 533, was posted on Wed Oct 29 '08 at 15:41:23
in reply to 83d707ec579-6876-817-00.htm

Re: Fatal error no. 0x00020070 in sub PACKWRIT on unit 8

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this simply means that you ran out of disk space.
The explanation of this error message is explicitly given
in F.A.Q, as well as in several previous posts to this forum.

Try direct SCF.

Regards,
Alex Granovsky


On Tue Oct 28 '08, Sibo wrote
-----------------------------
>Hi, I am new in Firefly--all I've successfully run is an optimization of 1-methylcyclohexene. Now I'm being a bit more ambitious and would like to optimize the structure of a poly-aromatic ligand with some copper ions. However, after determining the symmetries for the initial guess orbitals, my calculation crapped out. Can anyone tell me what this error means and how to fix it?



Message d4c0f196bHW-6877-946+00.htm, number 534, was posted on Wed Oct 29 '08 at 15:46:01
in reply to 829e844cbK2-6876-385-00.htm

Re: Selection of a basis set for Au111 surface

Alex Granovsky
gran@classic.chem.msu.su


Hi,

MINI basis set is not suitable for any practical calculations.
Use ECP basis sets. You can find some of them at e.g., http://www.emsl.pnl.gov/forms/basisform.html

Regards,
Alex Granovsky


On Tue Oct 28 '08, Mahesh Bhatt wrote
-------------------------------------
>Dear all,
>I have construced Au 111 surface by adding 13 Au atoms and then tried to calculate energy using B3LYP5 DFT parameter with basis sets STO-2G,3G,4G,5G,6G and 3-21G, 4-31G, 6-31G, 6-311G. But in PCGAMESS, there was a warning that illegal extended basis sets not requested and calculation terminated abnormally. In such a case, which basis set is available for Au surface in PCGAMESS? I have then used MINI basis set for calculation. But in case of a MINI basis set, there is zero charge transfer in Carbon and Nitrogen atoms when interacting an organic molecule with Au surface. This is a strange, isnt it?
>Is there any solution in PCGAMESS?
>Thanking you all
>Mahesh Bhatt
>University of Tsukuba
>Japan


Message 4df668afbHW-6879-1356+00.htm, number 535, was posted on Fri Oct 31 '08 at 22:36:00
in reply to 808be225Vl5-6876-961-00.htm

Re: about $EFIELD group

Alex Granovsky
gran@classic.chem.msu.su


Hi,

actually, the model is very simple and exact - external field
adds one-electron contribution to Hamiltonian, as well as changes
potential energy of nuclei.

The problem here is that modified Hamiltonian does not longer
have discrete spectrum at all! Formally, there are no bound
states anymore. However, as nobody uses really complete basis
sets but rather relatively compact basis sets designed to
describe bound states with square-integrable wavefunctions,
if you apply relatively weak electric field, you'll get
reasonable answer which will be in some sense good approximation
to the real physical picture. If you increase field strenght,
and especially if basis set has diffuse basis functions,
you'll see the "collapse" caused by the field - an attempt to
describe solution with electrons and nuclei infinitely separated.

Regards,
Alex Granovsky





On Tue Oct 28 '08, xuefei wrote
-------------------------------
>Dear all,

>Now, I am tring to use the $EFIELD group to study the influence of an external electric field on the molecule. I wonder what is the limit of the electric field strength? What theoretical model is used in calculations? Could you please intoduce some reference about that to me?

>Thanks a lot!

>Best wishes,
>xuefei


Message 0ccb39185au-6880-284-00.htm, number 536, was posted on Sat Nov 1 '08 at 04:44:02
values for convergence accelerators

Jonas Baltrusaitis
jonas-baltrusaitis@uiowa.edu


Hi,

what are the values enabled by default in convergence accelerators if using unrestricted calculations? I am talking about DAMP and SHIFT. Manual doesn't specify any


Message 775f60fd0y0-6880-309-00.htm, number 537, was posted on Sat Nov 1 '08 at 05:09:02
open shell SCF convergence for large silicon substituted benzene

arvey s. oliveros
yevra_16@yahoo.com


Good day sir/maam,

I am calculating the theoretical first vertical ionization potential of toluene molecules substituted with a bulky alkylsilane substituents, through quantum chemical calculation using PCGAMESS. The neutral molecules were optimize using HF/6-31G*. single point energy calculation was then done using Hartree Fock, B3LYP and MP2 method with a basis set of 6-31G**. For these calculations, my largest molecule has about 500 basis functions. i was using SCFTYP=RHF and SCFTYP=UHF for single point close shell and open shell energy calculations because i have limited time and limited computer resources to do higher SCF calculations. the basis is also selected for the same reason (also due to large size of molecules).
as of the moment im still doing the hartree fock calculations and later i will utilize its vector for DFT and MP2 calculation. For hartree fock method, I have no problem in SCF convergence for RHF (with GUESS=HUCKEL) but my UHF calculation gave me a lot of problem because it hardly converge. i have limited knowledge in PC GAMESS (I am still 19 year old BS CHEMISTRY student here in Philippines), so i would like to consult you regarding this matter....

1) is it logical to use molecular orbitals generated from STO-3G runs (with SCFTYP=UHF in hartree fock method) as a starting orbital for my hartree fock open shell calculations using the same method but at 6-31G**?

2) Is it alright to use the orbitals (for both open shell and close shell) from hartree fock calculations as a guess orbitals for B3LYP and MP2 single point energy calculation given that the basis set use is the same for the three calculations.

3) using only SCFTYP=UHF (for open shell) with the following inputs,
$SCF DAMP=.false. DEM=.false. DIIS=.false. DIRSCF=.true.
EXTRAP=.true. RSTRCT=.false. SHIFT=.false. SOSCF=.true.
what are your other recommendations for faster and efficient SCF convergence?

4) is the single point UHF calculation (UBLYP/6-31G** with GUESS=HUCKEL) that converge after long period (about 3 days and iteration of about more than 500) is a sign or indication of wrong MO'S???

THANKS A LOT SIR/MAAM FOR PATIENTLY ANSWERING MY QUESTIONS..... I HIGHLY APPRECIATE AND REGARD YOUR INTELLIGENCE IN COMPUTATIONAL CHEMISTRY. GOOD DAY AND MORE POWER TO YOUR WORKING TEAM!


Message cadbe2afNCe-6880-786-00.htm, number 538, was posted on Sat Nov 1 '08 at 13:07:23
ERROR of ILLEGAL BASIS FUNCTION TYPE

Fumihito Mohri
qzg00677@nifty.com


Dear Colleagues

Today, I performed ECP-computations for [Cu(H2O)4]2+, [Cu(H2O)5]2+ and
[Cu(H2O)6]2+ ,using PC-GAMESS ver.7.15 and the external basis function
of LANL2DZ. Among them, the computation for [Cu(H2O)4]2+ was successfully
made, but for the other two systems, the following error message appeared
. *** ERROR, ILLEGAL BASIS FUNCTION TYPE=@DZV IGAUSS= 0

I think that the number of atoms is limited in ECP-computation. Is this
correct? How should I do compute [Cu(H2O)5]2+ and [Cu(H2O)6]2+ ?

Best regards,
Fumihito Mohri

This message contains the 4 kb attachment
[ CuOH26-ECP.inp ] Cu(OH2)6-ECP


Message 589dd852bQn-6880-822+00.htm, number 539, was posted on Sat Nov 1 '08 at 13:41:55
in reply to 775f60fd0y0-6880-309-00.htm

Re: open shell SCF convergence for large silicon substituted benzene

Pedro Silva
pedros@ufp.pt


>1) is it logical to use molecular orbitals generated from STO-3G runs (with SCFTYP=UHF in hartree fock method) as a starting orbital for my hartree fock open shell calculations using the same method but at 6-31G**?


No... In PcGamess your input guess orbitals must be of the same basis set you want to converge.

>2) Is it alright to use the orbitals (for both open shell and close shell) from hartree fock calculations as a guess orbitals for B3LYP and MP2 single point energy calculation given that the basis set use is the same for the three calculations.

Yes

>3) using only SCFTYP=UHF (for open shell) with the following inputs,
>$SCF �DAMP=.false. DEM=.false. DIIS=.false. DIRSCF=.true. � � � � � � � � � �
> � � � � �EXTRAP=.true. RSTRCT=.false. SHIFT=.false. SOSCF=.true.
>what are your other recommendations for faster and efficient SCF convergence?

>4) is the single point UHF calculation (UBLYP/6-31G** with GUESS=HUCKEL) that converge after long period (about 3 days and iteration of about more than 500) is a sign or indication of wrong MO'S???

UHF convergence is most often helped by mixing the HOMO and LUMO guess orbitals. This is done with:

$guess mix=.t. $end


Message 4df668afbHW-6880-1103+00.htm, number 540, was posted on Sat Nov 1 '08 at 18:23:01
in reply to 589dd852bQn-6880-822+00.htm

Re^2: open shell SCF convergence for large silicon substituted benzene

Alex Granovsky
gran@classic.chem.msu.su


Hi,

On Sat Nov 1 '08, Pedro Silva wrote
-----------------------------------
>>>1) is it logical to use molecular orbitals generated from STO-3G runs (with SCFTYP=UHF in hartree fock method) as a starting orbital for my hartree fock open shell calculations using the same method but at 6-31G**?
>
>
>No... In PcGamess your input guess orbitals must be of the same basis set you want to converge.

The only exception is that you can use rdmini guess.

Regards,
Alex Granovsky


Message 4df668afbHW-6880-1108+00.htm, number 541, was posted on Sat Nov 1 '08 at 18:28:36
in reply to 775f60fd0y0-6880-309-00.htm

Re: open shell SCF convergence for large silicon substituted benzene

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>3) using only SCFTYP=UHF (for open shell) with the following inputs,
>$SCF DAMP=.false. DEM=.false. DIIS=.false. DIRSCF=.true.
> EXTRAP=.true. RSTRCT=.false. SHIFT=.false. SOSCF=.true.
>what are your other recommendations for faster and efficient SCF convergence?

The combination of convergence accelerators you selected
is usually not the best choice for UHF.

Try:

 $SCF DIRSCF=.true. DIIS=.true. ethrsh=1000 $end

If you are using PC GAMESS/Firefly v. 7.1.C
then you can try its improved SOSCF as well:

 $SCF DIRSCF=.true. DIIS=.false. soscf=.true. sogtol=1000 $end

Regards,
Alex Granovsky


Message 4df668afbHW-6880-1114+00.htm, number 542, was posted on Sat Nov 1 '08 at 18:33:41
in reply to cadbe2afNCe-6880-786-00.htm

Re: ERROR of ILLEGAL BASIS FUNCTION TYPE

Alex Granovsky
gran@classic.chem.msu.su


Dear Fumihito Mohri,

your input has illegal characters in it:

@DZV

instead of

 DZV

Regards,
Alex


Message 775f642c0y0-6881-295-00.htm, number 543, was posted on Sun Nov 2 '08 at 04:55:54
GUESS=Huckel versus GUESS=MOREAD in SCFTYP=UHF: gave different Ionization Energy Values

arvey s. oliveros
yevra_16@yahoo.com


i am using water molecule as a trial run for the inputs that i have made for my actual
molecules (which are large) so that i could see the variation of the total energy calculated
as influenced by my input. and hence how would my input affect the calculated ionization energy (IE) as well

Neutral Radical calculated experimental difference
(hartree) (hartree) IE (ev) IE (ev) kcal/mol
from RHF from UHF (calculated vs
experimental)
MP2_guess=huckel -76.1959582747 -75.7471185848 12.2135 12.6202 4.1885
MP2_Guess=moread & mix=.true. -76.1959582721 -75.6642291156 14.4691 12.6202 45.5183

The above results summarizes the ionization energy of water calculated at MP2/6-31G*. The RHF calculation gave identical results of total energy whether i use huckel as a guess or MO from HF calculation. but in the case of UHF, the calculated total energy of the frozen radical molecule differ by about 0.082889 hartree (2.25552447 eV).The huckel guess gave closer IE of the water compared to the guess=moread with MIX=.true. I believe that setting MIX=.true. possibly contribute to this significant error in calculating the total energy of the radical which in turn affected the ionization energy.

In relation to my alkylsilyl substituted toluene molecules (the actual molecules im dealing) i cannot utilize GUESS=HUCKEL because its SCFTYP=UHF did converge (i already tried). thus im only left with the GUESS=MOREAD (from hartree fock orbitals)and MIX=.true. which im hesitant to implement in my actual calculation due to the error on my preliminary calculations i made above. i tried setting MIX=.false. and the calculation terminated abnormally and it says "ERROR: PREMATURE END OF ORBITAL INPUT ENCOUNTERED LOOKING FOR ORBITAL 20 ELEMENT 1" i attach the output of my MP2_guess=moread with mix=.true. (SCFTYP=UHF) water calculation below for a detailed description. I did not include the other out's becuase there is no problem with those.

In this connection i would like to ask the following:

1. is there anyone who could give me a better input for SCFTYP=UHF with GUESS=MOREAD without setting MIX=.true.?

2. the manual says: MIX = rotate the alpha and beta HOMO and LUMO orbitals
so as to generate inequivalent alpha and beta
orbital spaces. This pertains to UHF singlets
only. This may require use of NOSYM=1 in $CONTRL
depending on your situation. (default=.FALSE.)
-> does it mean that MIX=.true. is only applicable for MULT=1 (singlets) in UHF calculation? and not in MULT=2 (which is the case for my molecules)?

3. the manual also says: For -UHF-, if NORB is not given, only the
occupied alpha and beta orbitals should be given, BACK TO
BACK. Otherwise, both alpha and beta orbitals must
consist of NORB vectors.

-> i dont understand what is meant by "BACK TO BACK" and how could i do this in my input?
-> i dont also understand what is meant by "Otherwise, both alpha and beta orbitals must
consist of NORB vectors" and how could i express this in my input files?

thanks for you answers

This message contains the 49 kb attachment
[ water_UMP2_moread_mixT_nosym0.out ] my UHF run with GUESS=MOREAD from HF MO's


Message cadbe2afNCe-6881-463+00.htm, number 544, was posted on Sun Nov 2 '08 at 07:43:09
in reply to 4df668afbHW-6880-1114+00.htm

Re^2: ERROR of ILLEGAL BASIS FUNCTION TYPE

Fumihito Mohri
qzg00677@nifty.com


Dear@Prof. Granovsky

Thank you for your answer. My problem has been resolved by
re-typing all 'DZV' in half size character.

Best regards,
Fumihito Mohri


Message 4df668afbHW-6881-823+00.htm, number 545, was posted on Sun Nov 2 '08 at 13:42:43
in reply to cadbe2afNCe-6881-463+00.htm

Re^3: ERROR of ILLEGAL BASIS FUNCTION TYPE

Alex Granovsky
gran@classic.chem.msu.su


Dear Prof. Fumihito Mohri,

it looks like the problem with extended character sets,
improper encoding, etc...

Note that PC GAMESS requires plain ASCII input file
containing Latin characters, numbers, and some special
symbols. Other symbols are allowed only in comments.

Best regards,
Alex Granovsky


On Sun Nov 2 '08, Fumihito Mohri wrote

>Thank you for your answer. My problem has been resolved by
>re-typing all 'DZV' in half size character.

>Best regards,
>Fumihito Mohri


Message db740624NCe-6881-958+00.htm, number 546, was posted on Sun Nov 2 '08 at 15:58:00
in reply to 4df668afbHW-6881-823+00.htm

Re^4: ERROR of ILLEGAL BASIS FUNCTION TYPE

Fumihito Mohri
qzg00677@nifty.com


Dear Prof. Granovsky
Thank you for your comment.
>it looks like the problem with extended character sets, improper
>encoding, etc...
Thatfs right.
During re-typing the all DZV characters, I found that DZV symbol of
5th oxygen atom has one-column space just before DZV characters, that is,
this DZV should be expressed not aseDZVf but as ' DZV'.@
This may be due to the fact that half-size setting in key-board
accidentally changed into full-size setting. We Japanese people
often encounter this problem, because Japanese letters are typed
in full-size setting, while English letters including plain ASCII
are typed in half-size setting.

Best regards,
Fumihito Mohri


Message 5b4e69f5Uee-6882-9-00.htm, number 547, was posted on Mon Nov 3 '08 at 00:10:55
Something strange with LANL2DZ

sanya
sanya@photonics.ru


Dear Alex,

I tried to use LANL2DZ ECP basis set for Iridium (from https://bse.pnl.gov/bse/portal) for some DFT (TDDFT) and RHF (CIS) calculations. To my surprise, the results were rather strange: bad SCF convergence, wrong energy (almost twice as low as it should be), and unphysical orbitals. Previously, I successfully used this basis set for the same molecule within Gaussian03, so I have data to compare with. I thoroughly checked the input but everything was the same as in my G03 runs (the number of electrons, basis functions, core electrons, and nuclear repulsion energy). Is it something wrong with the basis set? I tried to reproduce this calculation with GAMESS-US (with corresponding alterations in the input file), and it was OK. Attached is my input file.

This message contains the 44 kb attachment
[ IrPpy3_td_b3lyp.inp ] Input example with LANL2DZ basis set


Message 4df668afbHW-6882-19+00.htm, number 548, was posted on Mon Nov 3 '08 at 00:19:08
in reply to 0ccb39185au-6880-284-00.htm

Re: values for convergence accelerators

Alexei Popov
alexei.a.popov@gmail.com


Hi!

If you are interested in which particular options are ON by default,
you can find this information in output files I think.

Regards,
Alexei


On Sat Nov 1 '08, Jonas Baltrusaitis wrote
------------------------------------------
>Hi,

>what are the values enabled by default in convergence accelerators if using unrestricted calculations? I am talking about DAMP and SHIFT. Manual doesn't specify any


Message db740624NCe-6882-266+00.htm, number 549, was posted on Mon Nov 3 '08 at 04:26:01
in reply to 5b4e69f5Uee-6882-9-00.htm

Re: Something strange with LANL2DZ

Fumihito Mohri
qzg00677@nifty.com


Dear Sanya
I recommend you to try the followings.
(1)$GUESS GUESS=HCORE $END
Maybe, HUECKEL guess (default) is not appropriate to your Ir complex.
(2)changing symbols of none ECP atoms, for example,
C NONE -> C-ECP NONE
Please read $ECP description in input.pdf.

Best regards,
Fumihito Mohri


Message 5b4e69f5Uee-6882-278+00.htm, number 550, was posted on Mon Nov 3 '08 at 04:38:08
in reply to db740624NCe-6882-266+00.htm

Re^2: Something strange with LANL2DZ

sanya
sanya@photonics.ru


Dear Fumihito Mohri,

I've tried almost everything you recommend prior to bothering Dr. Granovsky. Unfortunately, it had no effect at all. At the same time, GAMESS-US works well with the same input (save for some minor changes concerning the difference in PCGAMESS and GAMESS-US input file format). Thank you, anyway.

Best regards,
Sanya

--------------------------------------
>Dear Sanya
>I recommend you to try the followings.
> (1)$GUESS GUESS=HCORE $END
> Maybe, HUECKEL guess (default) is not appropriate to your Ir complex.
> (2)changing symbols of none ECP atoms, for example,
> C NONE -> C-ECP NONE
> Please read $ECP description in input.pdf.

>Best regards,
>Fumihito Mohri


Message 4df668afbHW-6882-692+00.htm, number 551, was posted on Mon Nov 3 '08 at 11:32:59
in reply to 5b4e69f5Uee-6882-9-00.htm

Re: IMPORTANT - Something strange with LANL2DZ

Alex Granovsky
gran@classic.chem.msu.su


Dear all,

the origin of problems is very strange (if not really stupid)
format of basis sets which are available through
https://bse.pnl.gov/bse/portal (Basis set exchange portal)

E.g., for Carbon (taken from Sanya' input):

C      6.0      1.528682000        -0.802880000         1.067324000
S   9
  1   4233.0000000              0.0012200
  2    634.9000000              0.0093420
  3    146.1000000              0.0454520
  4     42.5000000              0.1546570
  5     14.1900000              0.3588660
  6      5.1480000              0.4386320
  7      1.9670000              0.1459180
  8      5.1480000              0.0000000
  9      0.4962000              0.0000000
S   9
  1   4233.0000000              0.0000000
  2    634.9000000              0.0000000
  3    146.1000000              0.0000000
  4     42.5000000              0.0000000
  5     14.1900000              0.0000000
  6      5.1480000              0.0000000
  7      1.9670000              0.0000000
  8      5.1480000             -0.1683670
  9      0.4962000              1.0600910

etc...

As you can see, there are primitives with exactly zero contraction
coefficients. This drives the gencon engine (which is turned on by
default) really mad, as it was written explicitly assuming that all
contraction coefficients are non-zero. I would like to stress that
this is quite a natural assumption as otherwise you'll just have
longer execution times (in the best case) without any benefits at
all. Moreover, gencon internally renormalizes contractions using
the smallest coefficient as the reference.

Thus, the problem is the incompatibility of gencon engine with the format of basis sets available through BSE portal.

As of now, there are two possible solutions:

1. Disable gencon engine ($contrl gencon=.f. $end)

or (the recommended one)

2. Manually reformat basis set provided by BSE portal to remove all zero contractions. E.g., for the fragment above correct basis would
be:

C      6.0      1.528682000        -0.802880000         1.067324000
S   7
  1   4233.0000000              0.0012200
  2    634.9000000              0.0093420
  3    146.1000000              0.0454520
  4     42.5000000              0.1546570
  5     14.1900000              0.3588660
  6      5.1480000              0.4386320
  7      1.9670000              0.1459180
S   2
  8      5.1480000             -0.1683670
  9      0.4962000              1.0600910

etc...

The fix for this problem has been developed today.
The fix simply eliminates all zero contractions at
the earliest stages of input file parsing. However,
this fix will not be available before the release of the
first scheduled update to PC GAMESS/Firefly v. 7.1.C
binaries.

Thus, if you have recently gotten strange results with PC GAMESS,
please! look carefully at your basis set and follow the instructions provided above.

Regards,
Alex Granovsky



On Mon Nov 3 '08, sanya wrote
-----------------------------
>Dear Alex,

>I tried to use LANL2DZ ECP basis set for Iridium (from https://bse.pnl.gov/bse/portal) for some DFT (TDDFT) and RHF (CIS) calculations. To my surprise, the results were rather strange: bad SCF convergence, wrong energy (almost twice as low as it should be), and unphysical orbitals. Previously, I successfully used this basis set for the same molecule within Gaussian03, so I have data to compare with. I thoroughly checked the input but everything was the same as in my G03 runs (the number of electrons, basis functions, core electrons, and nuclear repulsion energy). Is it something wrong with the basis set? I tried to reproduce this calculation with GAMESS-US (with corresponding alterations in the input file), and it was OK. Attached is my input file.


Message 4df668afbHW-6882-696+00.htm, number 551, was edited on Mon Nov 3 '08 at 11:36:50
and replaces message 4df668afbHW-6882-692+00.htm

Re: IMPORTANT - Something strange with LANL2DZ

Alex Granovsky
gran@classic.chem.msu.su


Dear all,

the origin of problems is very strange (if not really stupid)
format of basis sets which are available through
https://bse.pnl.gov/bse/portal (Basis set exchange portal)

E.g., for Carbon (taken from Sanya' input):

C      6.0      1.528682000        -0.802880000         1.067324000
S   9
  1   4233.0000000              0.0012200
  2    634.9000000              0.0093420
  3    146.1000000              0.0454520
  4     42.5000000              0.1546570
  5     14.1900000              0.3588660
  6      5.1480000              0.4386320
  7      1.9670000              0.1459180
  8      5.1480000              0.0000000
  9      0.4962000              0.0000000
S   9
  1   4233.0000000              0.0000000
  2    634.9000000              0.0000000
  3    146.1000000              0.0000000
  4     42.5000000              0.0000000
  5     14.1900000              0.0000000
  6      5.1480000              0.0000000
  7      1.9670000              0.0000000
  8      5.1480000             -0.1683670
  9      0.4962000              1.0600910

etc...

As you can see, there are primitives with exactly zero contraction
coefficients. This drives the gencon engine (which is turned on by
default) really mad, as it was written explicitly assuming that all
contraction coefficients are non-zero. I would like to stress that
this is quite a natural assumption as otherwise you'll just have
longer execution times (in the best case) without having any
benefits at all. Moreover, gencon internally renormalizes
contractions using the smallest coefficient as the reference.

Thus, the problem is the incompatibility of gencon engine with the format of basis sets available through BSE portal.

As of now, there are two possible solutions:

1. Disable gencon engine ($contrl gencon=.f. $end)

or (the recommended one)

2. Manually reformat basis set provided by BSE portal to remove
all zero contractions. E.g., for the fragment above correct basis
set description would be:

C      6.0      1.528682000        -0.802880000         1.067324000
S   7
  1   4233.0000000              0.0012200
  2    634.9000000              0.0093420
  3    146.1000000              0.0454520
  4     42.5000000              0.1546570
  5     14.1900000              0.3588660
  6      5.1480000              0.4386320
  7      1.9670000              0.1459180
S   2
  8      5.1480000             -0.1683670
  9      0.4962000              1.0600910

etc...

The fix for this problem has been already developed today.
The fix simply eliminates all zero contractions at
the earliest stages of input file parsing. However,
this fix will not be available before the release of the
first scheduled update to PC GAMESS/Firefly v. 7.1.C
binaries.

Thus, if you have recently gotten strange results with PC GAMESS,
please! look carefully at your basis set and follow the instructions provided above.

Regards,
Alex Granovsky



On Mon Nov 3 '08, sanya wrote
-----------------------------
>Dear Alex,

>I tried to use LANL2DZ ECP basis set for Iridium (from https://bse.pnl.gov/bse/portal) for some DFT (TDDFT) and RHF (CIS) calculations. To my surprise, the results were rather strange: bad SCF convergence, wrong energy (almost twice as low as it should be), and unphysical orbitals. Previously, I successfully used this basis set for the same molecule within Gaussian03, so I have data to compare with. I thoroughly checked the input but everything was the same as in my G03 runs (the number of electrons, basis functions, core electrons, and nuclear repulsion energy). Is it something wrong with the basis set? I tried to reproduce this calculation with GAMESS-US (with corresponding alterations in the input file), and it was OK. Attached is my input file.

[ This message was edited on Mon Nov 3 '08 by the author ]


Message 5b4e69c1Uee-6882-945+00.htm, number 552, was posted on Mon Nov 3 '08 at 15:45:13
in reply to 4df668afbHW-6882-696+00.htm

Re^2: IMPORTANT - Something strange with LANL2DZ

sanya
sanya@photonics.ru


Thank you very much! it really works!

Message c06b52111oc-6882-1057-00.htm, number 553, was posted on Mon Nov 3 '08 at 17:37:07
Torsion angle PES scan using DLC

Pasquale
morvillo@portici.enea.it


Hi,
I would like to make torsion angle PES scan using DLC.
I started with the example of the butadiene molecule reported in this site (bond length scan) and changed it to make a torsion around the atoms 1,2,3,4 but I got an error.
Can you help me?


$CONTRL SCFTYP=RHF RUNTYP=RSURFACE NZVAR=24 $END
$SYSTEM TIMLIM=100 MWORDS=10 $END
$BASIS GBASIS=STO NGAUSS=3 $END
$STATPT NSTEP=200 METHOD=GDIIS NOREG=5 $END
$SURF NDISP1=10 DISP1=10 IVEC1(1)=2,1 IGRP1(1)=1 ORIG1=0 REUSE=.T. $END
$ZMAT DLC=.T. AUTO=.T. AUTOFV=.T. $END
$ZMAT IFZMAT(1)=3,1,2,3,4 $END
$ZMAT NONVDW(1)=1,2 $END
$DATA
Butadiene
C1
C 6.0 -1.481450 -0.583385 0.188625
C 6.0 -0.723784 0.516553 0.061501
C 6.0 0.589576 0.530702 -0.220831
C 6.0 1.455228 -0.458012 0.049682
H 1.0 -1.084939 -1.603704 0.065897
H 1.0 -2.556345 -0.499052 0.419335
H 1.0 -1.225062 1.492569 0.187823
H 1.0 0.993524 1.437189 -0.705543
H 1.0 1.159982 -1.396368 0.545304
H 1.0 2.520570 -0.357575 -0.216207
$END

Message c317208dbQn-6883-877+00.htm, number 554, was posted on Tue Nov 4 '08 at 14:36:43
in reply to c06b52111oc-6882-1057-00.htm

Re: Torsion angle PES scan using DLC

Pedro Silva
pedros@ufp.pt


Dear Pasquale,

You should not use DLC coordinates for a torsion PES scan. You should explicitly define your z-matrix, as in the example below:


$scf dirscf=.t. damp=1 $end
$CONTRL ICHARG= 0 MULT= 1 MAXIT=100 �
SCFTYP=RHF RUNTYP=rsurface COORD=cart nzvar=51 $END �
$SYSTEM TIMLIM=10000 MEMORY=80000000 $END
$STATPT nstep=20 ifreez(1)=1 $END
$BASIS GBASIS=am1 $END

$GUESS GUESS=HUCKEL $END �
$surf ndisp1=12 disp1=15 vect1(1)=1 orig1=0 $end �
$ZMAT IZMAT(1)=
3 , 2 , 1 , 11 , 13 ,
1 , 1 , 2 , �
1 , 1 , 17 , �
1 , 1 , 11 , �
1 , 1 , 3 , �
1 , 2 , 4 , �
1 , 2 , 5 , �
1 , 3 , 16 , �
1 , 3 , 6 , �
1 , 3 , 15 , �
1 , 5 , 7 , �
1 , 5 , 9 , �
1 , 6 , 8 ,
1 , 6 , 9 ,
1 , 9 , 10 ,
1 , 11 , 13 ,
1 , 11 , 12 ,
1 , 13 , 14 ,
1 , 13 , 18 ,
1 , 18 , 19 ,
2 , 1 , 2 , 5 ,
2 , 5 , 9 , 6 ,
2 , 6 , 3 , 1 ,
2 , 1 , 2 , 4 ,
2 , 7 , 5 , 2 ,
2 , 10 , 9 , 5 ,
2 , 8 , 6 , 9 ,
2 , 15 , 3 , 6 ,
2 , 16 , 3 , 6 ,
2 , 3 , 1 , 17 ,
2 , 3 , 1 , 11 ,
2 , 1 , 11 , 12 ,
2 , 1 , 11 , 13 ,
2 , 11 , 13 , 14 ,
2 , 11 , 13 , 18 ,
2 , 13 , 18 , 19 ,
3 , 1 , 2 , 5 , 9 ,
3 , 5 , 9 , 6 , 3 ,
3 , 6 , 3 , 1 , 2 ,
3 , 3 , 1 , 2 , 4 ,
3 , 7 , 5 , 2 , 4 ,
3 , 10 , 9 , 5 , 7 ,
3 , 8 , 6 , 9 , 10 ,
3 , 15 , 3 , 6 , 8 ,
3 , 16 , 3 , 6 , 8 ,
3 , 17 , 1 , 3 , 16 ,
3 , 6 , 3 , 1 , 11 ,
3 , 3 , 1 , 11 , 12 ,
3 , 1 , 11 , 13 , 14 ,
3 , 1 , 11 , 13 , 18 ,
3 , 11 , 13 , 18 , 19
$end
$DATA
C8H10O
C1
CARBON �6.0 � -0.050386550 0.291143367 0.531215078
CARBON �6.0 � -0.482590538 -1.140864975 0.289158735
CARBON �6.0 � -0.940519643 1.263241995 -0.285324469
HYDROGEN 1.0 � �0.236086376 -1.939881665 0.449165874
CARBON �6.0 � -1.763741042 -1.437663337 0.014541057
CARBON �6.0 � -2.404847137 0.901838330 -0.219224428
HYDROGEN 1.0 �-2.077536804 -2.473953358 -0.083415771
HYDROGEN 1.0 �-3.145475790 1.687624576 -0.345494796
CARBON �6.0 � -2.774939669 -0.383843055 -0.106461337
HYDROGEN 1.0 � -3.824725359 -0.665139135 -0.110827734
CARBON �6.0 � �1.407969146 0.592464234 0.280994424
HYDROGEN �1.0 � 1.728060729 1.611229519 0.494807739
CARBON �6.0 � 2.342533065 -0.266512992 -0.127122834
HYDROGEN �1.0 � 2.122466841 -1.302711382 -0.373229052
HYDROGEN �1.0 � -0.772361699 2.288498555 0.062866240
HYDROGEN �1.0 � -0.615177395 1.240619495 -1.338299543
HYDROGEN �1.0 � -0.248450484 0.500834642 1.600061910
OXYGEN � �8.0 � 3.655285547 0.113714106 -0.263483251
HYDROGEN 1.0 � �4.148578450 -0.606278030 -0.672082493
$END


Pedro S


Message c317208dbQn-6883-884+00.htm, number 555, was posted on Tue Nov 4 '08 at 14:44:09
in reply to c317208dbQn-6883-877+00.htm

Re^2: Torsion angle PES scan using DLC

Pedro Silva
pedros@ufp.pt


Regarding my last email:
I have found that that during PES scans, the constrained torsion only remains constrained during the first 5 torsion scan steps. Afterwards, its value becomes unpredictable. This does not prevent one from running the PES scan, but it does force us to separate it in several smaller scans with ndisp1=5 each.

Message d4c0f196bHW-6883-1064+00.htm, number 556, was posted on Tue Nov 4 '08 at 17:45:00
in reply to c317208dbQn-6883-877+00.htm

Re^2: Torsion angle PES scan using DLC

Alex Granovsky
gran@classic.chem.msu.su


Dear all,

you CAN use DLCs for bonds, angles, and torsions scans with Firefly.
The complete documentation will be available from our site within
a week or so, meanwhile, I've attached a couple of samples in
the hope they will be useful. Please post your questions on these samples directly to this forum.

Regards,
Alex Granovsky



On Tue Nov 4 '08, Pedro Silva wrote
-----------------------------------
>Dear Pasquale,

>You should not use DLC coordinates for a torsion PES scan. You should explicitly define your z-matrix, as in the example below:

This message contains the 2 kb attachment
[ firefly_dlc_scans.rar ] Firefly sample files for scans in DLCs


Message d4c0f196bHW-6883-1067+00.htm, number 557, was posted on Tue Nov 4 '08 at 17:47:34
in reply to c317208dbQn-6883-884+00.htm

Re^3: Torsion angle PES scan using DLC

Alex Granovsky
gran@classic.chem.msu.su


Hi,

this is because the conversion from internals to Cartesians
did not converge at one of the intermediate points. Firefly
has better conversion routine so your sample works just fine
with it.

Regards,
Alex Granovsky


On Tue Nov 4 '08, Pedro Silva wrote
-----------------------------------
>Regarding my last email:
>I have found that that during PES scans, the constrained torsion only remains constrained during the first 5 torsion scan steps. Afterwards, its value becomes unpredictable. This does not prevent one from running the PES scan, but it does force us to separate it in several smaller scans with ndisp1=5 each.


Message 98585054SWH-6883-1187-00.htm, number 558, was posted on Tue Nov 4 '08 at 19:47:23
Eigenvalues

Manh-Thuong
manhthuong_z@yahoo.com


Hello everyone,

Can someone teach me how to print out all eigenvalues with higher precisions?

Thanks,
Manh



Message 5b50dbc51oc-6883-1404+00.htm, number 559, was posted on Tue Nov 4 '08 at 23:24:16
in reply to d4c0f196bHW-6883-1064+00.htm

Re^3: Torsion angle PES scan using DLC

Pasquale
morvillo@portici.enea.it


Thanks for the example files.
Regards

On Tue Nov 4 '08, Alex Granovsky wrote
--------------------------------------
>Dear all,

>you CAN use DLCs for bonds, angles, and torsions scans with Firefly.
>The complete documentation will be available from our site within
>a week or so, meanwhile, I've attached a couple of samples in
>the hope they will be useful. Please post your questions on these samples directly to this forum.

>Regards,
>Alex Granovsky
>
>
>
>On Tue Nov 4 '08, Pedro Silva wrote
>-----------------------------------
>>Dear Pasquale,

>>You should not use DLC coordinates for a torsion PES scan. You should explicitly define your z-matrix, as in the example below:

>


Message 4df668afbHW-6884-951+00.htm, number 560, was posted on Wed Nov 5 '08 at 15:50:44
in reply to 98585054SWH-6883-1187-00.htm

Re: Eigenvalues

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the only way is to change source code and recompile binaries.

Regards,
Alex Granovsky


On Tue Nov 4 '08, Manh-Thuong wrote
-----------------------------------
>Hello everyone,

>Can someone teach me how to print out all eigenvalues with higher precisions?

>Thanks,
>Manh

>
>


Message 4df668afbHW-6884-954+00.htm, number 561, was posted on Wed Nov 5 '08 at 15:53:41
in reply to 0ccb39185au-6880-284-00.htm

Re: values for convergence accelerators

Alex Granovsky
gran@classic.chem.msu.su


Hi,

the proper values are selected automatically.
Do you have some specific reason asking about these variables?

Regards,
Alex Granovsky


On Sat Nov 1 '08, Jonas Baltrusaitis wrote
------------------------------------------
>Hi,

>what are the values enabled by default in convergence accelerators if using unrestricted calculations? I am talking about DAMP and SHIFT. Manual doesn't specify any


Message dd080c8eaS1-6885-266-00.htm, number 562, was posted on Thu Nov 6 '08 at 04:26:31
A proble about triplet state optimization

Mark
zhanglm770@163.com


Dear all
I am encountering a problem when doing the triplet state optimization using CIS. PC GAMESS ends with a notification of
----------------------------------------------------------------
*** TOO MANY ITERATIONS IN AOCPCG *** MXCPIT= 50

MOST OFTEN THIS IS CAUSED BY USE OF AN INAPPROPRIATE WAVEFUNCTION.
CHANGE SCFTYP, OR CHECK HOMO/LUMO FILLING ORDER.

Even though I add the control keyword MXCPIT=1000 in $CPHF, the notification still shows up.
------------------------------------------------------------------
Here is the input part of my job
-------------------------------------------------------------------
$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE ICHARG=2 MULT=1
CITYP=CIS MAXIT=1000 ECP=SBKJC $END
$SYSTEM MEMORY=100000000 TIMLIM=525600 $END
$BASIS gbasis=sbkjc $END
$STATPT $END
$CPHF MXCPIT=1000 $END
$cis NSTATE=1 NDAVIT=1000 MULT=3 $END
$SCF DIRSCF=.TRUE. $END
$DATA
```````````````````````````````````
-----------------------------------------------

These are the outputs of my job

-----------------------------------------------------------
SOLVE THE COUPLED-PERTURBED HARTREE-FOCK EQUATIONS IN THE
ATOMIC-ORBITAL BASIS
-------------------------------------------------------------
ITER RESPONSE ERROR IMPROVED
1 ?.83356E-01 1
``````````````````````````````````````
50 1.27973E-04 1



*** TOO MANY ITERATIONS IN AOCPCG *** MXCPIT= 50


MOST OFTEN THIS IS CAUSED BY USE OF AN INAPPROPRIATE WAVEFUNCTION.
CHANGE SCFTYP, OR CHECK HOMO/LUMO FILLING ORDER.
-------------------------------------------------------------
Can anybody kindly help me with this problem?

Thank you in advance!

Have a nice day!


Message 5395f7047Tv-6885-792-00.htm, number 563, was posted on Thu Nov 6 '08 at 13:12:45
non-gradient triplet optimization problem

Denis Zavelev
denis.zavelev@gmail.com


Hello!

I'm trying to compute strange molecule HO4+ (HO2+ planar symmetric ion and O2 molecule that's perpendicular to the plane of the ion) with Firefly 7.1.C.
First I tried to compute this using single-CPU PC under MOPSLinux but it's appeared to be long...
So I decided to use my PC based on Intel CoreQuad Q6600 under Ubuntu 7.10.
I tried both statically and dynamically linked with MPICH versions but the following I got:
=======(statically linked)=====
TID 4242 caught signal 8, exiting.

Dump of registers follows

eax :: 0xffffffff, edx :: 0xffffffff
ecx :: 0x00000000, ebx :: 0x00000000
esi :: 0x097d80eb, edi :: 0x097dd1e7
ebp :: 0xf5d6d6aa, esp :: 0xf5d6d624
eip :: 0x1025314d, eflags :: 0x00210287

cs :: 0x0023
ds :: 0x0097
es :: 0x0097
ss :: 0x0097
fs :: 0x01cf
gs :: 0x0000

Stack backtrace

esp :: 0xf5d6d6aa, ebp :: 0x00000000, eip :: 0x00360000
===============================
=======(dynamically linked)====
TID 4288 caught signal 8, exiting.

Dump of registers follows

eax :: 0xffffffff, edx :: 0xffffffff
ecx :: 0x00000000, ebx :: 0x00000000
esi :: 0x097f3eeb, edi :: 0x097f8fe7
ebp :: 0xffa3cd2a, esp :: 0xffa3cca4
eip :: 0x1025314d, eflags :: 0x00210287

cs :: 0x0023
ds :: 0x002b
es :: 0x002b
ss :: 0x002b
fs :: 0x01cf
gs :: 0x0063

Stack backtrace

esp :: 0xffa3cd2a, ebp :: 0x00000000, eip :: 0x00360000
===============================

Here's my input file:
===============================
$CONTRL RUNTYP=TRUDGE MPLEVL=2 NOSYM=1 MAXIT=300 SCFTYP=UHF MULT=3 ICHARG=1
COORD=HINT $END
! $CONTRL EXETYP=CHECK $END
$SYSTEM TIMLIM=30000 MWORDS=320 $END
$BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NFFUNC=1 DIFFSP=.t. $END
$GUESS GUESS=HUCKEL PRTMO=.t. $END
$SCF DAMP=.t. SHIFT=.t. NCONV=6 DIRSCF=.t. FDIFF=.t. ETHRSH=0.05 SOSCF=.t. $END
$TRUDGE NPAR=10 IEX(1)=21,22,31,32,41,42,43,51,52,53 $END
$P2P P2P=.t. DLB=.t. XDLB=.t. $END
$SMP CSMTX=.t. CALL64=.t. $END
$DATA
HO4+
C1
HYDROGEN 1. LC 0.00 0.0 0.00 -
OXYGEN 8. PCC 1.2239615 122.07151 0.00 +
OXYGEN 8. PCC 1.2239615 57.31351 0.00 +
OXYGEN 8. NPCC 1.2239615 90.0 147.621 +
OXYGEN 8. NPCC 1.2239615 90.0 147.621 -
$END
===============================

CPU seems not to be overheated, so what's wrong?


Message d4c0f196bHW-6885-854+00.htm, number 564, was posted on Thu Nov 6 '08 at 14:14:08
in reply to 5395f7047Tv-6885-792-00.htm

Re: non-gradient triplet optimization problem

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Signal 8 could be the either floating point exception, or integer divide by zero exception.

Did you run this job in parallel? If yes, how many CPU cores was used?
What was the exact command line? Could you please attach output file if any?

Regards,
Alex Granovsky

On Thu Nov 6 '08 1:12pm, Denis Zavelev wrote
--------------------------------------------
>Hello!

>I'm trying to compute strange molecule HO4+ (HO2+ planar symmetric ion and O2 molecule that's perpendicular to the plane of the ion) with Firefly 7.1.C.
>First I tried to compute this using single-CPU PC under MOPSLinux but it's appeared to be long...
>So I decided to use my PC based on Intel CoreQuad Q6600 under Ubuntu 7.10.
>I tried both statically and dynamically linked with MPICH versions but the following I got:
>=======(statically linked)=====
>TID 4242 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097d80eb, edi :: 0x097dd1e7
>ebp :: 0xf5d6d6aa, esp :: 0xf5d6d624
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x0097
>es :: 0x0097
>ss :: 0x0097
>fs :: 0x01cf
>gs :: 0x0000

>Stack backtrace

>esp :: 0xf5d6d6aa, ebp :: 0x00000000, eip :: 0x00360000
>===============================
>=======(dynamically linked)====
>TID 4288 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097f3eeb, edi :: 0x097f8fe7
>ebp :: 0xffa3cd2a, esp :: 0xffa3cca4
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x0063

>Stack backtrace

>esp :: 0xffa3cd2a, ebp :: 0x00000000, eip :: 0x00360000
>===============================

>Here's my input file:
>===============================
> $CONTRL RUNTYP=TRUDGE MPLEVL=2 NOSYM=1 MAXIT=300 SCFTYP=UHF MULT=3 ICHARG=1
> COORD=HINT $END
>! $CONTRL EXETYP=CHECK $END
> $SYSTEM TIMLIM=30000 MWORDS=320 $END
> $BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NFFUNC=1 DIFFSP=.t. $END
> $GUESS GUESS=HUCKEL PRTMO=.t. $END
> $SCF DAMP=.t. SHIFT=.t. NCONV=6 DIRSCF=.t. FDIFF=.t. ETHRSH=0.05 SOSCF=.t. $END
> $TRUDGE NPAR=10 IEX(1)=21,22,31,32,41,42,43,51,52,53 $END
> $P2P P2P=.t. DLB=.t. XDLB=.t. $END
> $SMP CSMTX=.t. CALL64=.t. $END
> $DATA
> HO4+
>C1
> HYDROGEN 1. LC 0.00 0.0 0.00 -
> OXYGEN 8. PCC 1.2239615 122.07151 0.00 +
> OXYGEN 8. PCC 1.2239615 57.31351 0.00 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 -
> $END
>===============================

>CPU seems not to be overheated, so what's wrong?


Message d4c0f196bHW-6885-876+00.htm, number 564, was edited on Thu Nov 6 '08 at 14:35:50
and replaces message d4c0f196bHW-6885-854+00.htm

Re: non-gradient triplet optimization problem

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Signal 8 means this is either floating point exceptions, or integer divide by zero exception.

Did you run this job in parallel? If yes, how many CPU cores was used?
What was the exact command line? Could you please attach output file if any?

I believe this exception may be caused by some errors
in your input file (HINT cordinates are really tricky)
as you specified ten parameters (NPAR=10) to optimize,
while the number of internal coordinates is only 9 (3*5-6=9)

Regards,
Alex Granovsky


On Thu Nov 6 '08 1:12pm, Denis Zavelev wrote
--------------------------------------------
>Hello!

>I'm trying to compute strange molecule HO4+ (HO2+ planar symmetric ion and O2 molecule that's perpendicular to the plane of the ion) with Firefly 7.1.C.
>First I tried to compute this using single-CPU PC under MOPSLinux but it's appeared to be long...
>So I decided to use my PC based on Intel CoreQuad Q6600 under Ubuntu 7.10.
>I tried both statically and dynamically linked with MPICH versions but the following I got:
>=======(statically linked)=====
>TID 4242 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097d80eb, edi :: 0x097dd1e7
>ebp :: 0xf5d6d6aa, esp :: 0xf5d6d624
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x0097
>es :: 0x0097
>ss :: 0x0097
>fs :: 0x01cf
>gs :: 0x0000

>Stack backtrace

>esp :: 0xf5d6d6aa, ebp :: 0x00000000, eip :: 0x00360000
>===============================
>=======(dynamically linked)====
>TID 4288 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097f3eeb, edi :: 0x097f8fe7
>ebp :: 0xffa3cd2a, esp :: 0xffa3cca4
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x0063

>Stack backtrace

>esp :: 0xffa3cd2a, ebp :: 0x00000000, eip :: 0x00360000
>===============================

>Here's my input file:
>===============================
> $CONTRL RUNTYP=TRUDGE MPLEVL=2 NOSYM=1 MAXIT=300 SCFTYP=UHF MULT=3 ICHARG=1
> COORD=HINT $END
>! $CONTRL EXETYP=CHECK $END
> $SYSTEM TIMLIM=30000 MWORDS=320 $END
> $BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NFFUNC=1 DIFFSP=.t. $END
> $GUESS GUESS=HUCKEL PRTMO=.t. $END
> $SCF DAMP=.t. SHIFT=.t. NCONV=6 DIRSCF=.t. FDIFF=.t. ETHRSH=0.05 SOSCF=.t. $END
> $TRUDGE NPAR=10 IEX(1)=21,22,31,32,41,42,43,51,52,53 $END
> $P2P P2P=.t. DLB=.t. XDLB=.t. $END
> $SMP CSMTX=.t. CALL64=.t. $END
> $DATA
> HO4+
>C1
> HYDROGEN 1. LC 0.00 0.0 0.00 -
> OXYGEN 8. PCC 1.2239615 122.07151 0.00 +
> OXYGEN 8. PCC 1.2239615 57.31351 0.00 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 -
> $END
>===============================

>CPU seems not to be overheated, so what's wrong?

[ This message was edited on Thu Nov 6 '08 at 2:35pm by the author ]


Message d4c0f196bHW-6885-877+00.htm, number 564, was edited on Thu Nov 6 '08 at 14:37:10
and replaces message d4c0f196bHW-6885-876+00.htm

Re: non-gradient triplet optimization problem

Alex Granovsky
gran@classic.chem.msu.su


Hi,

Signal 8 means this is either floating point exceptions, or integer divide by zero exception error.

Did you run this job in parallel? If yes, how many CPU cores was used?
What was the exact command line? Could you please attach output file if any?

I believe this exception may be caused by some errors
in your input file (HINT cordinates are really tricky),
as you specified ten parameters (NPAR=10) to optimize,
while the number of internal coordinates is only 9 (3*5-6=9)

Regards,
Alex Granovsky


On Thu Nov 6 '08 1:12pm, Denis Zavelev wrote
--------------------------------------------
>Hello!

>I'm trying to compute strange molecule HO4+ (HO2+ planar symmetric ion and O2 molecule that's perpendicular to the plane of the ion) with Firefly 7.1.C.
>First I tried to compute this using single-CPU PC under MOPSLinux but it's appeared to be long...
>So I decided to use my PC based on Intel CoreQuad Q6600 under Ubuntu 7.10.
>I tried both statically and dynamically linked with MPICH versions but the following I got:
>=======(statically linked)=====
>TID 4242 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097d80eb, edi :: 0x097dd1e7
>ebp :: 0xf5d6d6aa, esp :: 0xf5d6d624
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x0097
>es :: 0x0097
>ss :: 0x0097
>fs :: 0x01cf
>gs :: 0x0000

>Stack backtrace

>esp :: 0xf5d6d6aa, ebp :: 0x00000000, eip :: 0x00360000
>===============================
>=======(dynamically linked)====
>TID 4288 caught signal 8, exiting.

>Dump of registers follows

>eax :: 0xffffffff, edx :: 0xffffffff
>ecx :: 0x00000000, ebx :: 0x00000000
>esi :: 0x097f3eeb, edi :: 0x097f8fe7
>ebp :: 0xffa3cd2a, esp :: 0xffa3cca4
>eip :: 0x1025314d, eflags :: 0x00210287

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x0063

>Stack backtrace

>esp :: 0xffa3cd2a, ebp :: 0x00000000, eip :: 0x00360000
>===============================

>Here's my input file:
>===============================
> $CONTRL RUNTYP=TRUDGE MPLEVL=2 NOSYM=1 MAXIT=300 SCFTYP=UHF MULT=3 ICHARG=1
> COORD=HINT $END
>! $CONTRL EXETYP=CHECK $END
> $SYSTEM TIMLIM=30000 MWORDS=320 $END
> $BASIS GBASIS=N31 NGAUSS=6 NDFUNC=2 NFFUNC=1 DIFFSP=.t. $END
> $GUESS GUESS=HUCKEL PRTMO=.t. $END
> $SCF DAMP=.t. SHIFT=.t. NCONV=6 DIRSCF=.t. FDIFF=.t. ETHRSH=0.05 SOSCF=.t. $END
> $TRUDGE NPAR=10 IEX(1)=21,22,31,32,41,42,43,51,52,53 $END
> $P2P P2P=.t. DLB=.t. XDLB=.t. $END
> $SMP CSMTX=.t. CALL64=.t. $END
> $DATA
> HO4+
>C1
> HYDROGEN 1. LC 0.00 0.0 0.00 -
> OXYGEN 8. PCC 1.2239615 122.07151 0.00 +
> OXYGEN 8. PCC 1.2239615 57.31351 0.00 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 +
> OXYGEN 8. NPCC 1.2239615 90.0 147.621 -
> $END
>===============================

>CPU seems not to be overheated, so what's wrong?

[ This message was edited on Thu Nov 6 '08 at 2:37pm by the author ]


Message 5395f7047Tv-6885-889+00.htm, number 565, was posted on Thu Nov 6 '08 at 14:48:42
in reply to d4c0f196bHW-6885-877+00.htm

Re^2: non-gradient triplet optimization problem

Denis Zavelev
denis.zavelev@gmail.com


Hello, Alex!

>Signal 8 means this is either floating point exceptions, or integer divide by zero exception error.
>Did you run this job in parallel? If yes, how many CPU cores was used?
Yes, 4 CPUs were used.

>What was the exact command line? Could you please attach output file if any?
./pcgamess -f -p -o $inpfile.out -ex . -t scratch -p4pg procgrp
(it was the command line from the script used)

>I believe this exception may be caused by some errors
>in your input file (HINT cordinates are really tricky),
>as you specified ten parameters (NPAR=10) to optimize,
>while the number of internal coordinates is only 9 (3*5-6=9)
If so, why this computation breaks after ~2 minutes while the same computation on single CPU PC still goes from yesterday's evening?


Message 5395f7047Tv-6885-889+00.htm, number 565, was edited on Thu Nov 6 '08 at 14:49:06
Re^2: non-gradient triplet optimization problem

Denis Zavelev
denis.zavelev@gmail.com


Hello, Alex!

>Signal 8 means this is either floating point exceptions, or integer divide by zero exception error.
>Did you run this job in parallel? If yes, how many CPU cores was used?
Yes, 4 CPU cores were used.

>What was the exact command line? Could you please attach output file if any?
./pcgamess -f -p -o $inpfile.out -ex . -t scratch -p4pg procgrp
(it was the command line from the script used)

>I believe this exception may be caused by some errors
>in your input file (HINT cordinates are really tricky),
>as you specified ten parameters (NPAR=10) to optimize,
>while the number of internal coordinates is only 9 (3*5-6=9)
If so, why this computation breaks after ~2 minutes while the same computation on single CPU PC still goes from yesterday's evening?

[ This message was edited on Thu Nov 6 '08 at 2:49pm by the author ]


Message 775f6ff50y0-6885-1016-00.htm, number 566, was posted on Thu Nov 6 '08 at 16:56:37
sample input file UHF calculation anyone?

arvey s. oliveros
yevra_16@yahoo.com


im a 19 yr old college, BS Chemistry student here in the Philippines and im working on my thesis right now
under the field of computational chemistry.

i had tried using orbitals generated from single point energy calculation at UHF/6-31G** as
the initial guess for an aromatic organosilicon compound for another single point energy calculation
at UB3LYP/6-31G** and UMP2/6-31G**. Unfortunately, as i follow the total energy and the energy change
as the iteration progresses, the total energy changes drastically (one of the iteration even gave positive
energy value!) and the change is about more than 100 hartrees. i am pretty sure that there is something
wrong in my input file. i tried to look at on the samples but i found nothing related to what i want to do.

can anyone give me a "template" input for calculating single point energy for SCFTYP=UHF at higher
calculation (as in MP2 and B3LYP) using GUESS=MOREAD from open shell hartree fock orbital? im pretty
sure that my error has something to do with the $CONTROL, $GUESS and $SCF....

thanks alot!


Message 775f6ff50y0-6885-1016+00.htm, number 566, was edited on Thu Nov 6 '08 at 17:11:52
and replaces message 775f6ff50y0-6885-1016-00.htm

sample input file UHF calculation anyone?

arvey s. oliveros
yevra_16@yahoo.com


im a 19 yr old college, BS Chemistry student here in the Philippines and im working on my thesis right now
under the field of computational chemistry.

i had tried using orbitals generated from single point energy calculation at UHF/6-31G** as
the initial guess for an aromatic organosilicon compound for another single point energy calculation
at UB3LYP/6-31G** and UMP2/6-31G**. Unfortunately, as i follow the total energy and the energy change
as the iteration progresses, the total energy changes drastically (one of the iteration even gave positive
energy value!) and the change is about more than 100 hartrees. i am pretty sure that there is something
wrong in my input file. i tried to look at on the samples but i found nothing related to what i want to do.

i had no problem in running it in close shell calculation. But in openshell calculation, it is some
kind of complicated because i have to consider the alpha and beta spins which i dont know how to take
it into account in the input file...so...

can anyone give me a "template" input for calculating single point energy for SCFTYP=UHF at higher
calculation (as in MP2 and B3LYP) using GUESS=MOREAD from open shell hartree fock orbital? im pretty
sure that my error has something to do with the $CONTROL, $GUESS and $SCF....

thanks alot!

[ This message was edited on Thu Nov 6 '08 at 5:11pm by the author ]


Message 5b4c45d5Uee-6885-1414+00.htm, number 567, was posted on Thu Nov 6 '08 at 23:34:46
in reply to 4df668afbHW-6882-696+00.htm

Re^2: IMPORTANT - Something strange with LANL2DZ

sanya
sanya@photonics.ru


By the way, I suddenly found that basis sets in this strange format are available only via https://bse.pnl.gov/bse/portal, whereas the old version of this site, http://www.emsl.pnl.gov/forms/basisform-orig.html, gives basis sets in the correct form, without zero contraction coefficients. Probably, the old version lacks some new basis sets or formats, so one may try both versions.

Message 829e844cbK2-6886-454-00.htm, number 568, was posted on Fri Nov 7 '08 at 07:34:15
Necessity of same basis function

Mahesh Bhatt
mbhatt54@yahoo.com


Dear all,
I have chosen MINI basis set with d and p orbital for Bathocuproine (BCP) adsobed on metal (111) surface and metals are Ca, Mg and Al. I have done this for minimum bound configuration for a particular structure. Once I got minimum bound configuration, I used 6-31G basis set for calculation of energy levels and molecular orbitals for same structure. The cause is that I could not get minimum bound configuration for a particular structure for above mentioned metals using 6-31G basis set even at a larger distance than found by B3LYP5/MINI(d,p). Is this possible or not? If possible, is it reliable or not?
I will have to submit a paper in International Research Papers. So I am confused here. I have got minimum bound configuration for Ca at 4.4 Angstron, Mg at 4.0 Angstron and Al at 4.4 Angstron. When I used 6-31G basis set for similar structure to get minimum bound configuration, I could not get around 5.0 Angstron or higher value.
Thanking all
Mahesh Bhatt
University of Tsukuba
Japan

Message 4df668afbHW-6886-764+00.htm, number 569, was posted on Fri Nov 7 '08 at 12:44:21
in reply to dd080c8eaS1-6885-266-00.htm

Re: A proble about triplet state optimization

Alex Granovsky
gran@classic.chem.msu.su


Hi,

As to

 $CPHF MXCPIT=... $END

option, it does not have any effect in the case of CIS (or MP2).

Instead, for CIS you can use the

 $CIS MAXITC=... $END

option.

However, most likely this will not help to solve convergence problem.

Sometimes, default conjugate gradient solver just becomes
numerically unstable. To avoid instabilities, you can
either increase icut (e.g., $contrl icut=10 $end),
or, alternatively, use

 $CIS CHFSLV=DIIS $END

which is slower but much more stable.


Regards,
Alex Granovsky


On Thu Nov 6 '08 4:26am, Mark wrote
-----------------------------------
>Dear all
>I am encountering a problem when doing the triplet state optimization using CIS. PC GAMESS ends with a notification of
>----------------------------------------------------------------
> *** TOO MANY ITERATIONS IN AOCPCG *** MXCPIT= 50

> MOST OFTEN THIS IS CAUSED BY USE OF AN INAPPROPRIATE WAVEFUNCTION.
> CHANGE SCFTYP, OR CHECK HOMO/LUMO FILLING ORDER.

>Even though I add the control keyword MXCPIT=1000 in $CPHF, the notification still shows up.
>------------------------------------------------------------------
>Here is the input part of my job
>-------------------------------------------------------------------
> $CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE ICHARG=2 MULT=1
> CITYP=CIS MAXIT=1000 ECP=SBKJC $END
> $SYSTEM MEMORY=100000000 TIMLIM=525600 $END
> $BASIS gbasis=sbkjc $END
> $STATPT $END
> $CPHF MXCPIT=1000 $END
> $cis NSTATE=1 NDAVIT=1000 MULT=3 $END
> $SCF DIRSCF=.TRUE. $END
> $DATA
>```````````````````````````````````
>-----------------------------------------------

>These are the outputs of my job

>-----------------------------------------------------------
>SOLVE THE COUPLED-PERTURBED HARTREE-FOCK EQUATIONS IN THE
> ATOMIC-ORBITAL BASIS
>-------------------------------------------------------------
>ITER RESPONSE ERROR IMPROVED
>1 ?.83356E-01 1
>``````````````````````````````````````
>50 1.27973E-04 1
>
>
>
> *** TOO MANY ITERATIONS IN AOCPCG *** MXCPIT= 50
>
>
> MOST OFTEN THIS IS CAUSED BY USE OF AN INAPPROPRIATE WAVEFUNCTION.
> CHANGE SCFTYP, OR CHECK HOMO/LUMO FILLING ORDER.
>-------------------------------------------------------------
>Can anybody kindly help me with this problem?

>Thank you in advance!

>Have a nice day!
>


Message 4df668afbHW-6886-880+00.htm, number 570, was posted on Fri Nov 7 '08 at 14:40:18
in reply to 5395f7047Tv-6885-889+00.htm

Re^3: non-gradient triplet optimization problem

Alex Granovsky
gran@classic.chem.msu.su


Hi,

indeed there is the funny bug in the old-style UMP2
energy code which does not allow parallel execution
to be performed correctly more than once per job
execution.

The workaround is just to use PC GAMESS/Firefly
specific code, which is not affected by this bug
and is faster.

Just add:

 $mp2 method=1 $end 

to your input.

I'd also suggest you to add

 $contrl inttyp=hondo $end

as well, as you use basis with diffuse basis functions.

However, closer examination of your system shows that

1. The single-determinant wavefunction does not describe this system correctly, and
2. There is strong spin-contamination already at SCF level

Taking this into account I'd suggest you to use CASSCF+MCQDPT2
rather than UHF+UMP2 approach.

Regards,
Alex Granovsky






On Thu Nov 6 '08 2:49pm, Denis Zavelev wrote
--------------------------------------------
>Hello, Alex!

>>Signal 8 means this is either floating point exceptions, or integer divide by zero exception error.
>>Did you run this job in parallel? If yes, how many CPU cores was used?
>Yes, 4 CPU cores were used.

>>What was the exact command line? Could you please attach output file if any?
>./pcgamess -f -p -o $inpfile.out -ex . -t scratch -p4pg procgrp
>(it was the command line from the script used)

>>I believe this exception may be caused by some errors
>>in your input file (HINT cordinates are really tricky),
>>as you specified ten parameters (NPAR=10) to optimize,
>>while the number of internal coordinates is only 9 (3*5-6=9)
>If so, why this computation breaks after ~2 minutes while the same computation on single CPU PC still goes from yesterday's evening?


Message 4df668afbHW-6886-1044+00.htm, number 571, was posted on Fri Nov 7 '08 at 17:23:43
in reply to 775f6ff50y0-6885-1016+00.htm

Re: sample input file UHF calculation anyone?

Alex Granovsky
gran@classic.chem.msu.su


Hi,

to be actually able to help you, it would be very helpful
to have your input file(s) attached to your post (please
compress it by any archiver before posting to forum!).

Regards,
Alex


On Thu Nov 6 '08 5:11pm, arvey s. oliveros wrote
------------------------------------------------
>im a 19 yr old college, BS Chemistry student here in the Philippines and im working on my thesis right now
>under the field of computational chemistry.

>i had tried using orbitals generated from single point energy calculation at UHF/6-31G** as
>the initial guess for an aromatic organosilicon compound for another single point energy calculation
>at UB3LYP/6-31G** and UMP2/6-31G**. Unfortunately, as i follow the total energy and the energy change
>as the iteration progresses, the total energy changes drastically (one of the iteration even gave positive
>energy value!) and the change is about more than 100 hartrees. i am pretty sure that there is something
>wrong in my input file. i tried to look at on the samples but i found nothing related to what i want to do.

>i had no problem in running it in close shell calculation. But in openshell calculation, it is some
>kind of complicated because i have to consider the alpha and beta spins which i dont know how to take
>it into account in the input file...so...
>
>can anyone give me a "template" input for calculating single point energy for SCFTYP=UHF at higher
>calculation (as in MP2 and B3LYP) using GUESS=MOREAD from open shell hartree fock orbital? im pretty
>sure that my error has something to do with the $CONTROL, $GUESS and $SCF....

>thanks alot!


Message 4df668afbHW-6886-1057+00.htm, number 572, was posted on Fri Nov 7 '08 at 17:37:01
in reply to 829e844cbK2-6886-454-00.htm

Re: Necessity of same basis function

Alex Granovsky
gran@classic.chem.msu.su


Hi,

as far as I remember you have already asked similar question a month ago or so.

Did you examine the sample calculations I posted that time to help you?

Actually, there are several problems with your system:

1. First of all, you need to use basis sets which are flexible enough,
otherwise you'll get meaningless answers.

2. Your metal surface should be properly terminated at all its
edges and borders. Otherwise, there MUST be serious convergence
problems, especially if basis set is not of MINI type!.
As MINI is just a toy basis, it avoids multiple convergence problems.
This does not mean that solution will readily converged for the same
chemical system while using extended basis sets!

Regards,
Alex Granovsky



On Fri Nov 7 '08 7:34am, Mahesh Bhatt wrote
-------------------------------------------
>Dear all,
>I have chosen MINI basis set with d and p orbital for Bathocuproine (BCP) adsobed on metal (111) surface and metals are Ca, Mg and Al. I have done this for minimum bound configuration for a particular structure. Once I got minimum bound configuration, I used 6-31G basis set for calculation of energy levels and molecular orbitals for same structure. The cause is that I could not get minimum bound configuration for a particular structure for above mentioned metals using 6-31G basis set even at a larger distance than found by B3LYP5/MINI(d,p). Is this possible or not? If possible, is it reliable or not?
>I will have to submit a paper in International Research Papers. So I am confused here. I have got minimum bound configuration for Ca at 4.4 Angstron, Mg at 4.0 Angstron and Al at 4.4 Angstron. When I used 6-31G basis set for similar structure to get minimum bound configuration, I could not get around 5.0 Angstron or higher value.
>Thanking all
>Mahesh Bhatt
>University of Tsukuba
>Japan


Message 5b4eb7cfUee-6887-287-00.htm, number 573, was posted on Sat Nov 8 '08 at 04:47:09
PCM

sanya
sanya@photonics.ru


Dear Alex,

What is the model of PCM implemented in PCGAMESS? It is not the same as in GAMESS-US or G03.

In addition, it exhibits strange behavior: for one molecule PCMAnalyze procedure finishes after several steps, and for an almost similar molecule (just a different conformer or a protonated form) PCMAnalyze makes many hundreds of iterations. And it works so slowly that I've already put the idea of geometry optimization in PCM out of my head :( I need only CIS or TDDFT spectra.


Message 589dd852bQn-6887-937+00.htm, number 574, was posted on Sat Nov 8 '08 at 15:37:24
in reply to 5b4eb7cfUee-6887-287-00.htm

Re: PCM

Pedro Silva
pedros@ufp.pt


On Sat Nov 8 '08 4:47am, sanya wrote
------------------------------------
>Dear Alex,

>What is the model of PCM implemented in PCGAMESS? It is not the same as in GAMESS-US or G03.

That model is described in the post:

http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C3d4763b02bHW-6085-1169+00.htm

>In addition, it exhibits strange behavior: for one molecule PCMAnalyze procedure finishes after several steps, and for an almost similar molecule (just a different conformer or a protonated form) PCMAnalyze makes many hundreds of iterations. And it works so slowly that I've already put the idea of geometry optimization in PCM out of my head :( I need only CIS or TDDFT spectra.

You can disable PCManalyze with $pcm analyz=0 $end

Pedro


Message 5b4ecb43Uee-6887-1030+00.htm, number 575, was posted on Sat Nov 8 '08 at 17:10:27
in reply to 589dd852bQn-6887-937+00.htm

Re^2: PCM

sanya
sanya@photonics.ru


>You can disable PCManalyze with $pcm analyz=0 $end

Thank you, that's great! but will it affect the energies and spectra?


Message 589dd852bQn-6888-70+00.htm, number 576, was posted on Sun Nov 9 '08 at 01:10:06
in reply to 5b4ecb43Uee-6887-1030+00.htm

Re^3: PCM

Pedro Silva
pedros@ufp.pt


On Sat Nov 8 '08 5:10pm, sanya wrote
------------------------------------
>>You can disable PCManalyze with $pcm analyz=0 $end

>Thank you, that's great! but will it affect the energies and spectra?

In my experience, it does not change energies by more than 0.1 kcal.mol-1


Message db7404d6NCe-6888-804-00.htm, number 577, was posted on Sun Nov 9 '08 at 13:25:35
Geometry optimization for lowest-excited- singlet sates by DFT

Fumihito Mohri
qzg00677@nifty.com


Dear PC-GAMESS users,

Recently, I started computations on organic electro-luminescence materials.
Now, I encounter the problem how to do geometry optimization for lowest-
excited- singlet sates by DFT. To my understanding, geometry optimization
for singlet-ground state (S0), lowest-excited-singlet sate (S1) and
lowest-excited triplet state (T1) can be made as follows.

In Hartree-Fock method
S0: $CONTRL SCFTYP=RHF MULT=1 RUNTYP=OPTIMIZE

S1: $CONTRL SCFTYP=RHF MULT=1 CITYP=CIS RUNTYP=OPTIMIZE ...
$CIS MULT=1....

T1: (a) $CONTRL SCFTYP=RHF MULT=1 CITYP=CIS RUNTYP=OPTIMIZE ...
$CIS MULT=3....
(b) $CONTRL SCFTYP=UHF MULT=3 RUNTYP=OPTIMIZE ...
Note that (a) and (b) give different results in general.

In DFT method (ex. B3LYP)
S0: $CONTRL SCFTYP=RHF MULT=1 DFTTYP=B3LYP RUNTYP=OPTIMIZE

S1: How to do geometry optimization ?

T1: $CONTRL SCFTYP=UHF MULT=3 DFTTYP=B3LYP RUNTYP=OPTIMIZE

Best regatds,
Fumihito Mohri


Message 4df668afbHW-6888-812+00.htm, number 578, was posted on Sun Nov 9 '08 at 13:32:56
in reply to 5b4eb7cfUee-6887-287-00.htm

Re: PCM

Alex Granovsky
gran@classic.chem.msu.su


Dear Sanya,

In addition to points which Pedro already mentioned (thanks!),
I would like to make some general comments regarding your
(and perhaps other users) questions. I think this is important
as some things are either poorly or not documeted at all.

1. The only PCM model implemented in PC GAMESS is D-PCM. This is
because the PC GAMESS was forked out from GAMESS (US) at the time
when it was the only PCM model available in GAMESS (US) as well.

2. There are multiple internal problems specific to this solvent model, namely:

I think in the future we'll completely replace D-PCM model in the
PC GAMESS/Firefly by the fast, efficient, and stable implementation
of COSMO model.

Regards,
Alex Granovsky


On Sat Nov 8 '08 4:47am, sanya wrote
------------------------------------
>Dear Alex,

>What is the model of PCM implemented in PCGAMESS? It is not the same as in GAMESS-US or G03.

>In addition, it exhibits strange behavior: for one molecule PCMAnalyze procedure finishes after several steps, and for an almost similar molecule (just a different conformer or a protonated form) PCMAnalyze makes many hundreds of iterations. And it works so slowly that I've already put the idea of geometry optimization in PCM out of my head :( I need only CIS or TDDFT spectra.


Message db7404d5NCe-6888-948+00.htm, number 579, was posted on Sun Nov 9 '08 at 15:48:23
in reply to db7404d6NCe-6888-804-00.htm

Re: Geometry optimization for lowest-excited- singlet sates by DFT

Fumihito Mohri
qzg00677@nifty.com


p.s.
My problem has been resolved by 'Introduction to PC GAMESS/Firefly by Marek Checinskif.
According to the Frank-Condon principle, S1 geometry can be regarded as the same as
S0 geometry. Hence, input dat of S1 calculation is as follows.

$CONTRL SCFTYP=RHF (MULT=1) CITYP=TDDFT DFTTYP=B3LYP RUNTYP=ENERGY....
$TDDFT (MULT=1) NSTATE=.....
....
$DATA
Atomic coordinates obtained from the previous geometry optimization of S0

Regards,
Fumihito Mohri


Message 5b4eb508Uee-6888-1268+00.htm, number 580, was posted on Sun Nov 9 '08 at 21:07:48
in reply to 4df668afbHW-6888-812+00.htm

Re^2: PCM

sanya
sanya@photonics.ru


Thank you all!

Message 5b4eb508Uee-6888-1277+00.htm, number 581, was posted on Sun Nov 9 '08 at 21:18:20
in reply to db7404d5NCe-6888-948+00.htm

Re^2: Geometry optimization for lowest-excited- singlet sates by DFT

sanya
sanya@photonics.ru


It's not quite true. For absorption, S1 geometry is exactly the same as S0 due to Franck--Condon principle. But not for luminescence. After excitation (S0->S1 or S0->Sn), the molecule relaxes to the lowest vibrational level of S1 (possibly, from Sn) simultaneously with the structural relaxation. That is, you have to optimize the geometry of S1. Unfortunately, you can do this only by CIS in PCGAMESS or try to use other programs.

Message d4c0ee0abHW-6889-2+00.htm, number 582, was posted on Mon Nov 10 '08 at 00:02:31
in reply to 5b4eb508Uee-6888-1277+00.htm

Re^3: Geometry optimization for lowest-excited- singlet sates by DFT

Alex Granovsky
gran@classic.chem.msu.su


Another approach you can try for excited states
geometry optimization with PC GAMESS/Firefly is MCSCF.
State-specific gradients are available for both single-state
and state-averaged MCSCF, and both for GUGA and ALDET CI code.
And MCSCF is blazingly fast with PC GAMESS/Firefly!

Regards,
Alex Granovsky

On Sun Nov 9 '08 9:18pm, sanya wrote
------------------------------------
>It's not quite true. For absorption, S1 geometry is exactly the same as S0 due to Franck--Condon principle. But not for luminescence. After excitation (S0->S1 or S0->Sn), the molecule relaxes to the lowest vibrational level of S1 (possibly, from Sn) simultaneously with the structural relaxation. That is, you have to optimize the geometry of S1. Unfortunately, you can do this only by CIS in PCGAMESS or try to use other programs.


Message c0ace279U2I-6889-781-00.htm, number 583, was posted on Mon Nov 10 '08 at 13:00:42
Firefly (PCG71C) compatibility with Tinker 4.2

Veinardi Suendo
vsuendo@chem.itb.ac.id


Dear Sirs,
How to do QM/MM calculation with PCGAMESS/Tinker. Is there any recent manual or guideline? Can I use Firefly (PCG71C) with Tinker 4.2? I have found an old manual for PCG65 that works only with Tinker 3.7(http://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htm). Or is there any other MM package for doing QM/MM calculation with Firefly.

Thank you very much in advance,


Yours Sincerely,


Veinardi Suendo


Message c06b52111oc-6890-904-00.htm, number 585, was posted on Tue Nov 11 '08 at 15:03:58
PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


Hi,
I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:

Segmentation fault (core dumped)

What can I do?

I had no problem under Linux 32bits.

Regards


Message 5395f7047Tv-6890-937+00.htm, number 586, was posted on Tue Nov 11 '08 at 15:36:58
in reply to 4df668afbHW-6886-880+00.htm

Re^4: non-gradient triplet optimization problem

Denis Zavelev
denis.zavelev@gmail.com


Hello, Alex!

>indeed there is the funny bug in the old-style UMP2
>energy code which does not allow parallel execution
>to be performed correctly more than once per job
>execution.
>The workaround is just to use PC GAMESS/Firefly
>specific code, which is not affected by this bug
>and is faster.
>Just add:
>

 
 $mp2 method=1 $end 

>to your input.
Thanks!

>I'd also suggest you to add
>

 $contrl inttyp=hondo $end

>as well, as you use basis with diffuse basis functions.
Exactly! I don't know why did I forget this...

>However, closer examination of your system shows that
>1. The single-determinant wavefunction does not describe this system correctly, and
>2. There is strong spin-contamination already at SCF level

>Taking this into account I'd suggest you to use CASSCF+MCQDPT2
>rather than UHF+UMP2 approach.
I should use SCFTYP=MCSCF for the geometry optimization, right?


>>>I believe this exception may be caused by some errors
>>>in your input file (HINT cordinates are really tricky),
>>>as you specified ten parameters (NPAR=10) to optimize,
>>>while the number of internal coordinates is only 9 (3*5-6=9)
>>If so, why this computation breaks after ~2 minutes while the same computation on single CPU PC still goes from yesterday's evening?
BTW, that computation also crashed (after about a day)...
So it seems that I really put a mistake in $DATA section.

But I can't understand these HINT coordinates (a really stupid beginner, oh yeah!).
I found that article and read it thoroughly but it's real hard to find the difference between NPCC and CCPA type connections (as they are shown strictly the same!), furthermore, I can't understand what is "beta" angle in CCPA case. HINT coordinates have some common with well-known spherical coordinates, but seem to be the perverted version of the last ones. I couldn't find anything in the web that can help me to understand...

In this case it seems we have the following (using C1 symmetry just in case, though Cs will not solve the problem):
we are putting H atom in the centre and draw I,J,K (and 3 zeros for H atom, LC, "-").
Then we place O2 onto x axis, so we have LC "-" connection with H.
Placing O3 onto xz plane, we have PCC, "-", with bond length and angle.
But how to place the rest 2 oxygen atoms?
I suppose I should use CCPA, but what is beta?
Just in case: cartesian coordinates go below.
HYDROGEN 1.0 0.000000000 0.000000000 0.000000000
OXYGEN 8.0 1.223961499 0.000000000 0.000000000
OXYGEN 8.0 0.521949141 0.000000000 -1.107091165
OXYGEN 8.0 -0.875913455 0.655452553 0.548852733
OXYGEN 8.0 -0.875913455 -0.655452553 0.548852733


Message 5395f7047Tv-6890-949+00.htm, number 587, was posted on Tue Nov 11 '08 at 15:50:22
in reply to cac90ba9So7-6889-1218-00.htm

Re: Error in wrong TRUDGE type computation

Denis Zavelev
denis.zavelev@gmail.com


Hello!

Are you sure that your $DATA input gives you the right geometry? And are you sure that initial guess P(1)=2.5 is good?

> When i begin a TRUDGE type computation like following:
> $CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
> COORD=HINT ECP=SBKJC $END
> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
> $BASIS GBASIS=SBKJC $END
> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
> $TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
> $TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
> $STATPT OPTTOL=0.0001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
> $END

>i got a very terrible answer, unexpectedly, the enery is -7041.*****,it can not be that!
>i do not konw what is warong in my input file, can you help me ?


Message 5395f7047Tv-6890-958+00.htm, number 588, was posted on Tue Nov 11 '08 at 15:58:13
in reply to c06b52111oc-6890-904-00.htm

Re: PC GAMESS under Linux Ubuntu 64bits

Denis Zavelev
denis.zavelev@gmail.com


Hello!

>I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:
>Segmentation fault (core dumped)
>What can I do?
>I had no problem under Linux 32bits.
Check your CPU cores temperature as it may overheat.
You may use sensors package to see CPU temperature (under load!) and xosview to see cores' load.


Message dcd109c1NCe-6890-983+00.htm, number 589, was posted on Tue Nov 11 '08 at 16:23:19
in reply to 5b4eb508Uee-6888-1277+00.htm

Re^3: Geometry optimization for lowest-excited- singlet sates by DFT

Fumihito Mohri
qzg00677@nifty.com


Dear Sanya

Thank you for your kind comment.
Especially, the followings is helpful for me.
>That is, you have to optimize the geometry of S1.
> Unfortunately, you can do this only by CIS in PCGAMESS
Now, I use CIS for the optimization of S1.

Regards,
Fumihito Mohri


Message dcd109c1NCe-6890-987+00.htm, number 590, was posted on Tue Nov 11 '08 at 16:27:23
in reply to d4c0ee0abHW-6889-2+00.htm

Re^4: Geometry optimization for lowest-excited- singlet sates by DFT

Fumihito Mohri
qzg00677@nifty.com


Dear Prof. Granovsky
Thank you for your helpful advice.

>Another approach you can try for excited states
>geometry optimization with PC GAMESS/Firefly is MCSCF.
>State-specific gradients are available for both single-state
>and state-averaged MCSCF, and both for GUGA and ALDET CI code.
>And MCSCF is blazingly fast with PC GAMESS/Firefly!

Regards,
Mohri, Fumihito


Message 5b4eb645Uee-6890-1077+00.htm, number 591, was posted on Tue Nov 11 '08 at 17:56:49
in reply to dcd109c1NCe-6890-983+00.htm

Re^4: Geometry optimization for lowest-excited- singlet sates by DFT

sanya
sanya@photonics.ru


>Now, I use CIS for the optimization of S1.

It is very important to make sure that the state you are optimizing is qualitatively correct (e.g., the orbitals participating the transition have nonzero spatial overlap). In TDDFT, the correct excited state is not necessarily the lowest one, while in CIS it usually is. But be careful about root flipping and have in mind that potential energy surface in TDDFT qualitatively differs from that in CIS and both are far from reality ;)

Anyway, when optimizing excited states, you are close to the wind. So be extremely careful and don't take anything on trust :)

In fact, Alex is right, and MSCSF methods are much more reliable, at least, in a qualitative sense. But I'm afraid you won't be able to use them for your electroluminescent compounds. You may configure a MCSCF run with exetyp=check and see the memory and disk requirements. What if they are quite moderate?


Message 4df668afbHW-6890-1104+00.htm, number 592, was posted on Tue Nov 11 '08 at 18:23:53
in reply to c06b52111oc-6890-904-00.htm

Re: PC GAMESS under Linux Ubuntu 64bits

Alex Granovsky
gran@classic.chem.msu.su


Hi,

could you please provide more information?
It would be helpful to know:

1. The OS version and kernel version

2. The PC GAMESS/Firefly version details:

a) target processor
b) build number
c) MPI type
d) statically or dynamically linked

3. The exact comand line you used to launch PC GAMESS/Firefly

Regards,
Alex Granovsky


On Tue Nov 11 '08 3:03pm, Pasquale wrote
----------------------------------------
>Hi,
>I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:

>Segmentation fault (core dumped)

>What can I do?

>I had no problem under Linux 32bits.

>Regards


Message 4df668afbHW-6890-1106+00.htm, number 593, was posted on Tue Nov 11 '08 at 18:26:27
in reply to cac90ba9So7-6889-1218-00.htm

Re: Error in wrong TRUDGE type computation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

could you please repost your input file?

There are some problems reading your coordinates
as well as atomic charge.

It would be fine if you simply attach your input to the post.

Regards,
Alex Granovsky


On Mon Nov 10 '08, ganymede wrote
---------------------------------
>Dear Prof. and all
> When i begin a TRUDGE type computation like following:

> $CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
> COORD=HINT ECP=SBKJC $END
> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
> $BASIS GBASIS=SBKJC $END
> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
> $TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
> $TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
> $STATPT OPTTOL=0.0001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
> $END

>i got a very terrible answer, unexpectedly, the enery is -7041.*****,it can not be that!
>i do not konw what is warong in my input file, can you help me ?

> Thank you very much!


Message 4df668afbHW-6890-1109+00.htm, number 594, was posted on Tue Nov 11 '08 at 18:29:44
in reply to c0ace279U2I-6889-781-00.htm

Re: Firefly (PCG71C) compatibility with Tinker 4.2

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I've asked James Kress to answer your question.

Regards,
Alex Granovsky

On Mon Nov 10 '08 1:00pm, Veinardi Suendo wrote
-----------------------------------------------
>Dear Sirs,
>How to do QM/MM calculation with PCGAMESS/Tinker. Is there any recent manual or guideline? Can I use Firefly (PCG71C) with Tinker 4.2? I have found an old manual for PCG65 that works only with Tinker 3.7(http://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htmhttp://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htm). Or is there any other MM package for doing QM/MM calculation with Firefly.

>Thank you very much in advance,
>
>
>Yours Sincerely,
>
>
>Veinardi Suendo


Message 4df668afbHW-6890-1116-00.htm, number 595, was posted on Tue Nov 11 '08 at 18:36:23
Announce of beta-test program of PC GAMESS/Firefly for Mac/OSX

Alex Granovsky
gran@classic.chem.msu.su


Dear all,

everybody who is interested to participate in beta-test program
for PC GAMESS/Firefly running under Mac/OSX, please contact me by email for details.

Regards,
Alex Granovsky


Message cac90ba9So7-6890-1181+00.htm, number 584, was edited on Tue Nov 11 '08 at 19:41:26
and replaces message cac90ba9So7-6889-1218-00.htm

Error in wrong TRUDGE type computation

ganymede
ganymede1314520@gmail.com


Dear Prof. and all
When i begin a TRUDGE type computation like following:

$CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
COORD=HINT ECP=SBKJC $END
$SYSTEM TIMLIM=525600 MEMORY=1000000 $END
$BASIS GBASIS=SBKJC $END
$SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
$TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
$TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
$STATPT OPTTOL=0.0001 NSTEP=200 $END
$DATA
Title
DNH 8

Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
$END

i got a very terrible answer, unexpectedly, the enery is -7041.*****,it can not be that!
i do not konw what is warong in my input file, can you help me ?

Thank you very much!

[ This message was edited on Tue Nov 11 '08 at 7:41pm by the author ]


Message cac90ba9So7-6890-1183+00.htm, number 584, was edited on Tue Nov 11 '08 at 19:42:47
and replaces message cac90ba9So7-6890-1181+00.htm

Error in wrong TRUDGE type computation

ganymede
ganymede1314520@gmail.com


Dear Prof. and all
When i begin a TRUDGE type computation like following:

$CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
COORD=HINT ECP=SBKJC $END
$SYSTEM TIMLIM=525600 MEMORY=1000000 $END
$BASIS GBASIS=SBKJC $END
$SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
$TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
$TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
$STATPT OPTTOL=0.0001 NSTEP=200 $END
$DATA
Title
DNH 8

Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
$END

i got a very terrible answer, unexpectedly, the enery is -7041.*****,it can not be that!
i do not konw what is warong in my input file, can you help me ?

Thank you very much!

[ This message was edited on Tue Nov 11 '08 at 7:42pm by the author ]


Message dbf63a10So7-6890-1184+00.htm, number 596, was posted on Tue Nov 11 '08 at 19:44:47
in reply to 4df668afbHW-6890-1106+00.htm

Re^2: Error in wrong TRUDGE type computation

ganymede
ganymede1314520@gmail.com


here is the input file:

$CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
COORD=HINT ECP=SBKJC $END
$SYSTEM TIMLIM=525600 MEMORY=1000000 $END
$BASIS GBASIS=SBKJC $END
$SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
$TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
$TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
$STATPT OPTTOL=0.0001 NSTEP=200 $END
$DATA
Title
DNH 8

Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
$END

This message contains the 1 kb attachment
[ Fe2.inp ] Fe.inp


Message 4df668afbHW-6890-1194+00.htm, number 592, was edited on Tue Nov 11 '08 at 19:54:08
and replaces message 4df668afbHW-6890-1104+00.htm

Re: PC GAMESS under Linux Ubuntu 64bits

Alex Granovsky
gran@classic.chem.msu.su


Hi,

could you please provide some more information?
It would be very helpful to know:

1. The OS version and Linux kernel version

2. The details of hardware configuration

3. The PC GAMESS/Firefly version details:

a) target processor
b) The PC GAMESS/Firefly build number
c) MPI type
d) Is this statically or dynamically linked version?

4. The exact command line you used to launch PC GAMESS/Firefly

Regards,
Alex Granovsky


On Tue Nov 11 '08 3:03pm, Pasquale wrote
----------------------------------------
>Hi,
>I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:

>Segmentation fault (core dumped)

>What can I do?

>I had no problem under Linux 32bits.

>Regards

[ This message was edited on Tue Nov 11 '08 at 7:54pm by the author ]


Message 4df668afbHW-6890-1200+00.htm, number 597, was posted on Tue Nov 11 '08 at 20:00:33
in reply to 5b4eb645Uee-6890-1077+00.htm

Re^5: Geometry optimization for lowest-excited- singlet sates by DFT

Alex Granovsky
gran@classic.chem.msu.su


>In fact, Alex is right, and MSCSF methods are much more reliable, at least, in a qualitative sense. But I'm afraid you won't be able to use them for your electroluminescent compounds. You may configure a MCSCF run with exetyp=check and see the memory and disk requirements. What if they are quite moderate?

Just one more comment - if indeed MCSCF is applicable to your systems,
please read carefully the PC GAMESS/Firefly specific documentation on MCSCF,
as the use of proper settings is very important for really speedy calculations.

Regards,
Alex


Message d4c0f196bHW-6890-1238+00.htm, number 598, was posted on Tue Nov 11 '08 at 20:38:30
in reply to 5395f7047Tv-6890-937+00.htm

Re^5: non-gradient triplet optimization problem

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>I should use SCFTYP=MCSCF for the geometry optimization, right?
Yes, you have to specify scftyp=mcscf and mplevl=2

The proper selection of active space is very important. Look at UHF NOs first, examine localized ROHF orbitals, CIS natural orbitals, etc...

>BTW, that computation also crashed (after about a day)...
>So it seems that I really put a mistake in $DATA section.

What was the error message?


>But I can't understand these HINT coordinates (a really stupid beginner, oh yeah!).
>I found that article and read it thoroughly but it's real hard to find the difference between NPCC and CCPA type connections (as they are shown strictly the same!), furthermore, I can't understand what is "beta" angle in CCPA case. HINT coordinates have some common with well-known spherical coordinates, but seem to be the perverted version of the last ones. I couldn't find anything in the web that can help me to understand...

>In this case it seems we have the following (using C1 symmetry just in case, though Cs will not solve the problem):
>we are putting H atom in the centre and draw I,J,K (and 3 zeros for H atom, LC, "-").
>Then we place O2 onto x axis, so we have LC "-" connection with H.
>Placing O3 onto xz plane, we have PCC, "-", with bond length and angle.
>But how to place the rest 2 oxygen atoms?
>I suppose I should use CCPA, but what is beta?
>Just in case: cartesian coordinates go below.
>HYDROGEN 1.0 0.000000000 0.000000000 0.000000000
>OXYGEN 8.0 1.223961499 0.000000000 0.000000000
>OXYGEN 8.0 0.521949141 0.000000000 -1.107091165
>OXYGEN 8.0 -0.875913455 0.655452553 0.548852733
>OXYGEN 8.0 -0.875913455 -0.655452553 0.548852733

Frankly, I'm not a specialist in HINT coordinates at all.
However, I can look into source code to find out why there are
formally ten parameters and how to eliminate the unused one.

Regards,
Alex



Message d4c0ee0abHW-6890-1370+00.htm, number 599, was posted on Tue Nov 11 '08 at 22:51:05
in reply to dbf63a10So7-6890-1184+00.htm

Re^3: Error in wrong TRUDGE type computation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

there are several errors in your original input file,
which result in completely wrong starting geometry and answer.
Remember that HINT coordinates are really tricky!

I've fixed your input and checked that it works normally.
I've also included the sample of gradient optimization of your system -
note that for DFT, you do not need to use trudge for geometry optimization.

Regards,
Alex





On Tue Nov 11 '08 7:44pm, ganymede wrote
----------------------------------------
>here is the input file:

> $CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
> COORD=HINT ECP=SBKJC $END
> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
> $BASIS GBASIS=SBKJC $END
> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
> $TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
> $TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
> $STATPT OPTTOL=0.0001 NSTEP=200 $END
> $DATA
>Title
>DNH 8

>Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
> $END
>

This message contains the 48 kb attachment
[ fe2_fixed.rar ] Fixed Fe2 inputs and outputs


Message 5b50ff0d1oc-6890-1391+00.htm, number 600, was posted on Tue Nov 11 '08 at 23:11:16
in reply to 4df668afbHW-6890-1194+00.htm

Re^2: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


Hi,

>could you please provide some more information?
>It would be very helpful to know:

>1. The OS version and Linux kernel version

Ubuntu 8.10 (Intrepid Ibex) 64-bits

>2. The details of hardware configuration

single PC with Intel Core 2 Duo E6550

>3. The PC GAMESS/Firefly version details:

PC GAMESS/Firefly version 7.1.C

>a) target processor

Pentium 4, Pentium D, Xeon, Intel Core 2

>b) The PC GAMESS/Firefly build number

build number 5014

>c) MPI type
>d) Is this statically or dynamically linked version?

Linux MPICH fully statically linked

>4. The exact command line you used to launch PC GAMESS/Firefly

./pcgamess -o demo.out dir1 dir2 -p4pg procgrp

where the file procgrp contains the text -> local 2

Using the same syntax, everything was fine under linux 32 bits (Puppy Linux 4.0)

Regards

>Regards,
>Alex Granovsky
>
>
>On Tue Nov 11 '08 3:03pm, Pasquale wrote
>----------------------------------------
>>Hi,
>>I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:

>>Segmentation fault (core dumped)

>>What can I do?

>>I had no problem under Linux 32bits.

>>Regards


Message 53e132e71oc-6891-20+00.htm, number 601, was posted on Wed Nov 12 '08 at 00:20:12
in reply to 5395f7047Tv-6890-958+00.htm

Re^2: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


On Tue Nov 11 '08 3:58pm, Denis Zavelev wrote
---------------------------------------------
>Hello!

>>I tried to run pc gamess under Linux Ubuntu 64bits (Live CD) and I got the error:
>>Segmentation fault (core dumped)
>>What can I do?
>>I had no problem under Linux 32bits.
>Check your CPU cores temperature as it may overheat.
>You may use sensors package to see CPU temperature (under load!) and xosview to see cores' load.

The CPU temperature is fine.


Message d4c0ee0abHW-6891-95+00.htm, number 602, was posted on Wed Nov 12 '08 at 01:35:10
in reply to 5b50ff0d1oc-6890-1391+00.htm

Re^3: PC GAMESS under Linux Ubuntu 64bits

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>>1. The OS version and Linux kernel version
>Ubuntu 8.10 (Intrepid Ibex) 64-bits

What's the kernel version ("uname -a" output)?

>>2. The details of hardware configuration
>single PC with Intel Core 2 Duo E6550
That's just not enough information...

What is the contents of /etc/fstab and what are the filesystems/mount points etc...

>>3. The PC GAMESS/Firefly version details:
>PC GAMESS/Firefly version 7.1.C
>Pentium 4, Pentium D, Xeon, Intel Core 2
>build number 5014
>Linux MPICH fully statically linked
>>4. The exact command line you used to launch PC GAMESS/Firefly
>./pcgamess -o demo.out dir1 dir2 -p4pg procgrp

The checklist to do:

1. Check if you can run it sequentially (without p4pg switch and removing procgrp file)
2. Check if the -nompi command line switch helps to run it sequentially
3. Check if dynamically linked version works normally.

>where the file procgrp contains the text -> local 2

Why are you using "local 2"?
Normaly on two cores you just need "local 1"...

Regards,
Alex Granovsky


Message cac90ba9So7-6891-719+00.htm, number 603, was posted on Wed Nov 12 '08 at 11:59:00
in reply to d4c0ee0abHW-6890-1370+00.htm

Re^4: Error in wrong TRUDGE type computation

ganymede
ganymede1314520@gmail.com


Dear Prof. Alex
Thank you very much!
But i have a question about XC functionals, why always pure functionals are much more difficult to converge than hybrid functionals?



On Tue Nov 11 '08 10:51pm, Alex Granovsky wrote
-----------------------------------------------
>Hi,

>there are several errors in your original input file,
>which result in completely wrong starting geometry and answer.
>Remember that HINT coordinates are really tricky!

>I've fixed your input and checked that it works normally. ?>I've also included the sample of gradient optimization of your system -
>note that for DFT, you do not need to use trudge for geometry optimization.

>Regards,
>Alex
>
>
>
>
>
>On Tue Nov 11 '08 7:44pm, ganymede wrote
>----------------------------------------
>>here is the input file:

>> $CONTRL SCFTYP=UHF RUNTYP=TRUDGE DFTTYP=b3lyp MAXIT=200 MULT=7
>> COORD=HINT ECP=SBKJC $END
>> $SYSTEM TIMLIM=525600 MEMORY=1000000 $END
>> $BASIS GBASIS=SBKJC $END
>> $SCF DIRSCF=.TRUE. FDIFF=.FALSE. $END
>> $TRUDGE OPTMIZ=GEOMETRY NPAR=1 IEX(1)=11 P(1)=2.5 $END
>> $TRURST TOLF=0.0001 TOLR=0.001 FNOISE=0.00001 $END
>> $STATPT OPTTOL=0.0001 NSTEP=200 $END
>> $DATA
>>Title
>>DNH 8

>>Fe ??6.0 LC ?0.00000 ??0.00000 ??.00000 + K
>> $END
>>


Message 5395f7047Tv-6891-908+00.htm, number 604, was posted on Wed Nov 12 '08 at 15:08:17
in reply to d4c0f196bHW-6890-1238+00.htm

Re^6: non-gradient triplet optimization problem

Denis Zavelev
denis.zavelev@gmail.com


Hello, Alex!

>>BTW, that computation also crashed (after about a day)...
>>So it seems that I really put a mistake in $DATA section.
>What was the error message?
*ERR* LI-02 absolute value of argument to arcsine, arccosine must not exceed one

>
>>But I can't understand these HINT coordinates (a really stupid beginner, oh yeah!).
>>I found that article and read it thoroughly but it's real hard to find the difference between NPCC and CCPA type connections (as they are shown strictly the same!), furthermore, I can't understand what is "beta" angle in CCPA case. HINT coordinates have some common with well-known spherical coordinates, but seem to be the perverted version of the last ones. I couldn't find anything in the web that can help me to understand...
>>In this case it seems we have the following (using C1 symmetry just in case, though Cs will not solve the problem):
>>we are putting H atom in the centre and draw I,J,K (and 3 zeros for H atom, LC, "-").
>>Then we place O2 onto x axis, so we have LC "-" connection with H.
>>Placing O3 onto xz plane, we have PCC, "-", with bond length and angle.
>>But how to place the rest 2 oxygen atoms?
>>I suppose I should use CCPA, but what is beta?
>>Just in case: cartesian coordinates go below.
>>HYDROGEN 1.0 0.000000000 0.000000000 0.000000000
>>OXYGEN 8.0 1.223961499 0.000000000 0.000000000
>>OXYGEN 8.0 0.521949141 0.000000000 -1.107091165
>>OXYGEN 8.0 -0.875913455 0.655452553 0.548852733
>>OXYGEN 8.0 -0.875913455 -0.655452553 0.548852733
>

>Frankly, I'm not a specialist in HINT coordinates at all.
It seems that Hilderbrandt is the only one specialist...

>However, I can look into source code to find out why there are
>formally ten parameters and how to eliminate the unused one.
Perhaps 2nd atom must be connected using LC instead of PCC, this will give totally 9 parameters. The only problem is to understand how to give 4th and 5 th atoms in CCPA, as attempt to use NPCC for them is being failed.


Message c06b52111oc-6891-1086+00.htm, number 605, was posted on Wed Nov 12 '08 at 18:05:49
in reply to d4c0ee0abHW-6891-95+00.htm

Re^4: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


Hi

>>>1. The OS version and Linux kernel version
>>Ubuntu 8.10 (Intrepid Ibex) 64-bits

>What's the kernel version ("uname -a" output)?

Linux ubuntu 2.6.27-7-generic #1 SMP Fri Oct 24 06:40:41 UTC 2008 x86_64 GNU/Linux


>>>2. The details of hardware configuration
>>single PC with Intel Core 2 Duo E6550
>That's just not enough information...

>What is the contents of /etc/fstab and what are the filesystems/mount points etc...

# /etc/fstab: static file system information.
#
#
proc /proc proc defaults 0 0
/host/ubuntu/disks/root.disk / ext3 loop,errors=remount-ro 0 1
/host/ubuntu/disks/boot /boot none bind 0 0
/host/ubuntu/disks/swap.disk none swap loop,sw 0 0

>>>3. The PC GAMESS/Firefly version details:
>>PC GAMESS/Firefly version 7.1.C
>>Pentium 4, Pentium D, Xeon, Intel Core 2
>>build number 5014
>>Linux MPICH fully statically linked
>>>4. The exact command line you used to launch PC GAMESS/Firefly
>>./pcgamess -o demo.out dir1 dir2 -p4pg procgrp

>The checklist to do:

>1. Check if you can run it sequentially (without p4pg switch and removing procgrp file)

I tried. I got the same error: Segmentation fault

>2. Check if the -nompi command line switch helps to run it sequentially

no.

>3. Check if dynamically linked version works normally.

I have to try

>>where the file procgrp contains the text -> local 2

>Why are you using "local 2"?
>Normaly on two cores you just need "local 1"...

I am using local 1. The same error.

>Regards,
>Alex Granovsky

Regards


Message c06b52111oc-6891-1088+00.htm, number 606, was posted on Wed Nov 12 '08 at 18:08:08
in reply to d4c0ee0abHW-6891-95+00.htm

Re^4: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


and under Ubuntu 32 bits, pcgamess works fine.

Message c3ce4a82XLp-6891-1158-00.htm, number 607, was posted on Wed Nov 12 '08 at 19:26:27
PCGAMESS and RESP CHARGES

Patrick
ps14@soton.ac.uk


Hi PCGAMESS Forum,

I would like to generate RESP charges for some small molecules using
PCGAMESS as the generator of the ESP points that are the input to the RESP
program. I've found instructions for doing this with GAMESS in the distribution
of AMBER9. I've got the latest version of PCGAMESS and I've assumed that it supports
this capability. Here are the relevant instructions for setting up the GAMESS input
that I used:
****************************************************************************
*******
1) Generate an appropiate GAMESS *.dat file which includes both the esp
points and the atomic coordinates. The GAMESS input file which does
this is shown here:

$CONTRL SCFTYP=RHF EXETYP=RUN RUNTYP=OPTIMIZE COORD=UNIQUE $END
$CONTRL MOLPLT=.TRUE. $END
$STATPT NSTEP=100 $END
$BASIS GBASIS=N31 NGAUSS=6 NDFUNC=1 $END
$ELPOT IEPOT=1 WHERE=PDC OUTPUT=PUNCH $END
$PDC PTSEL=CONNOLLY $END
$GUESS GUESS=HUCKEL $END
$DATA
eg180.pdb
C1
C01 6.0 -0.6388649947 0.4052483921 0.0000000000
...
...
H10 1.0 2.5277794177 0.0322897353 0.0000000000
$END

The runs completed successfully but did not produce the set of expected esp charges. Has anyone experienced similar issues? Is someone out there that has actually generated RESP charges using PCGAMESS to get esp points?

I am grateful for any hint regarding this issue.

Patrick


Message 4df668afbHW-6891-1410+00.htm, number 608, was posted on Wed Nov 12 '08 at 23:30:20
in reply to c06b52111oc-6891-1086+00.htm

Re^5: PC GAMESS under Linux Ubuntu 64bits

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>>3. Check if dynamically linked version works normally.
>I have to try

Please try and let us know.

Another thing to do is to run Firefly under strace,
e.g. strace ./pcgamess -nompi >out 2 >&1

This will produce (large) file with very helpful information.
Please send it back to me, not to the list!

And finally, you can run dynamically-linked version under gdb to see where exactly it crashes.

Regards,
Alex


Message 4ad81824Jpb-6892-70-00.htm, number 609, was posted on Thu Nov 13 '08 at 01:10:05
manual

Alex
deckuofm@yahoo.com


Dear PC GAMESS users,
Could you please specify link to download manual

Thank you in advance


Message 55b9038f49j-6892-548-00.htm, number 610, was posted on Thu Nov 13 '08 at 09:08:58
geometry optimization with a end-to-end distance constrain

jaber jahannbin Sardroodi
jsardroodi@azaruniv.edu


Dear pcgamess users,
Can anyone please help me for geometrical optimization of a curved configuration of tetracene (four benzen rings fused together) as its end-to-end distance be fixed? I need to find minimum energy of curved tetracene. Than you all

Message cbb1f7a17v5-6892-567-00.htm, number 611, was posted on Thu Nov 13 '08 at 09:27:45
URGENT! Please help!!!

James Phil
philx1012@yahoo.com


Newbie here so please bear with me...

My name is James Phil Flores and I'm currently working on the study using PCGAMESS. I'm finding the Bond Dissociation Energy of Selected Silicon-Containing compounds. I had found a basis set suitable for my compounds and my computed values are then compared to literature. When the dissociated atom is the H-radical (H.), the results are great. However, for calculations using other radicals (e.g CH3., SiH3., etc), there seem to be a great deviation with the literature values. With this, I would like to know if anyone could give any recommendation to make my study more reliable and to have better results. Should I increase my basis set? Should I change my method for calculation? Or is it really expected that my method and basis set used is good only for those calculation where the dissociated atom is the H radical? As of now, my chosen basis set and method is B3LYP/6-31+G(3d,p,f).

Also, I would like to know if anyone have any literature values for the following compounds (The H atom is the one being dissociated):
1)SiH3O-H
2)CH2=SiH-H
3)SiH3SiHCH2-H
4)SiH3CH2CH2-H
5)SiH3SiH2SiH2-H

If anyone doesn't know any literature values, can someone please suggest any book or sites that I could used to find the said values? Please do help me with this problem. It would really be a great help. Thank you very much!


Message c06b52111oc-6892-724+00.htm, number 612, was posted on Thu Nov 13 '08 at 12:03:38
in reply to 4df668afbHW-6891-1410+00.htm

Re^6: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


On Wed Nov 12 '08 11:30pm, Alex Granovsky wrote
-----------------------------------------------
>Hi,

>>>3. Check if dynamically linked version works normally.
>>I have to try

>Please try and let us know.

>Another thing to do is to run Firefly under strace,
>e.g. strace ./pcgamess -nompi >out 2 >&1

>This will produce (large) file with very helpful information.
>Please send it back to me, not to the list!

OK. I did.


>And finally, you can run dynamically-linked version under gdb to see where exactly it crashes.

how can I do this?

regards


Message c06b52111oc-6892-804+00.htm, number 612, was edited on Thu Nov 13 '08 at 13:24:32
and replaces message c06b52111oc-6892-724+00.htm

Re^6: PC GAMESS under Linux Ubuntu 64bits

Pasquale
morvillo@portici.enea.it


On Wed Nov 12 '08 11:30pm, Alex Granovsky wrote
-----------------------------------------------
>Hi,

>>>3. Check if dynamically linked version works normally.
>>I have to try

>Please try and let us know.

>Another thing to do is to run Firefly under strace,
>e.g. strace ./pcgamess -nompi >out 2 >&1

>This will produce (large) file with very helpful information.
>Please send it back to me, not to the list!

OK. I did.


>And finally, you can run dynamically-linked version under gdb to see where exactly it crashes.

I tried using the version for
Linux MPICH (using ssh as remote shell by default), dynamically linked NPTL-based version optimized for Pentium 4, Pentium D, Xeon, Intel Core 2

and it works fine!!!! I was able to carry out the calculations succesfully.

Regards

regards

[ This message was edited on Thu Nov 13 '08 at 1:24pm by the author ]


Message c06b52111oc-6892-894+00.htm, number 612, was edited on Thu Nov 13 '08 at 14:54:01
and replaces message c06b52111oc-6892-804+00.htm

Re^6: PC GAMESS under Linux Ubuntu 64bits: dynamically linked version works normally

Pasquale
morvillo@portici.enea.it


On Wed Nov 12 '08 11:30pm, Alex Granovsky wrote
-----------------------------------------------
>Hi,

>>>3. Check if dynamically linked version works normally.
>>I have to try

>Please try and let us know.

>Another thing to do is to run Firefly under strace,
>e.g. strace ./pcgamess -nompi >out 2 >&1

>This will produce (large) file with very helpful information.
>Please send it back to me, not to the list!

OK. I did.


>And finally, you can run dynamically-linked version under gdb to see where exactly it crashes.

I tried using the version for
Linux MPICH (using ssh as remote shell by default), dynamically linked NPTL-based version optimized for Pentium 4, Pentium D, Xeon, Intel Core 2

and it works fine!!!! I was able to carry out the calculations succesfully.

Regards

regards

[ This message was edited on Thu Nov 13 '08 at 2:54pm by the author ]


Message c317208dbQn-6892-1120+00.htm, number 613, was posted on Thu Nov 13 '08 at 18:40:38
in reply to cbb1f7a17v5-6892-567-00.htm

Re: URGENT! Please help!!!

Pedro Silva
pedros@ufp.pt


>However, for calculations using other radicals (e.g CH3., SiH3., >etc), there seem to be a great deviation with the literature >values. With this, I would like to know if anyone could give any >recommendation to make my study more reliable and to have better >results.


Are you sure the bond is breaking homolytically when you remove .CH3, etc? You might try to get orbitals from the .CH3 radical and from your separate X. radical, and combine them using the "combo" program (available elsewhere in this site), and use this as your guess orbitals.

>Should I increase my basis set? Should I change my method for >calculation? Or is it really expected that my method and basis set >used is good only for those calculation where the dissociated atom B3LYP/6-31+G(3d,p,f).


Be sure to perform the computations as scftyp=UHF




>
>Also, I would like to know if anyone have any literature values for the following compounds (The H atom is the one being dissociated):
>1)SiH3O-H
>2)CH2=SiH-H
>3)SiH3SiHCH2-H
>4)SiH3CH2CH2-H
>5)SiH3SiH2SiH2-H
>
>If anyone doesn't know any literature values, can someone please suggest any book or sites that I could used to find the said values?

Try NIST Chem Webbook (webbook.nist.gov/chemistry )


Message 4df668afbHW-6892-1355+00.htm, number 614, was posted on Thu Nov 13 '08 at 22:35:38
in reply to 55b9038f49j-6892-548-00.htm

Re: geometry optimization with a end-to-end distance constrain

Alex Granovsky
gran@classic.chem.msu.su


Hi,

if you are using cartesian coordinates for geometry optimization,
the simplest way is just to freeze coordinates of two atoms on the
opposite sides of molecule. E.g.,

 $statpt ifreez(1)=1,2,3,7,8,9 $end

would freeze x,y,and z coordinates of first and third atom.

These numbers (1,2,3,7,8,9) are just the indices of elements
of the following sequence: x1,y1,z1,x2,y2,z2,x3,y2,z3, etc...

If you are using internals, you can freeze particular
bond distance either using ifreez array or (for DLCs)
ifzmat array of $zmat group.

Regards,
Alex Granovsky

On Thu Nov 13 '08 9:08am, jaber jahannbin Sardroodi wrote
---------------------------------------------------------
>Dear pcgamess users,
>Can anyone please help me for geometrical optimization of a curved configuration of tetracene (four benzen rings fused together) as its end-to-end distance be fixed? I need to find minimum energy of curved tetracene. Than you all


Message 4df668afbHW-6892-1357+00.htm, number 615, was posted on Thu Nov 13 '08 at 22:36:53
in reply to 4ad81824Jpb-6892-70-00.htm

Re: manual

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I'd suggest you to visit "Manuals" section
of the PC GAMESS/Firefly homepage.

Regards,
Alex Granovsky

On Thu Nov 13 '08 1:10am, Alex wrote
------------------------------------
>Dear PC GAMESS users,
>Could you please specify link to download manual

>Thank you in advance
>


Message d4c0ee0abHW-6892-1371+00.htm, number 616, was posted on Thu Nov 13 '08 at 22:51:21
in reply to c06b52111oc-6892-894+00.htm

Re^7: PC GAMESS under Linux Ubuntu 64bits: dynamically linked version works normally

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I checked the output of strace you sent to me and it seems that the problem is with the


sendmsg()

and

recvmsg()

calls - they are improperly handled by the 32-bit subsystem of your 64-bit Linux version.
You may file bug report to developers if you like :-)

Regards,
Alex Granovsky


Message d4c0ee0abHW-6892-1380+00.htm, number 617, was posted on Thu Nov 13 '08 at 22:59:49
in reply to cac90ba9So7-6891-719+00.htm

Re^5: Error in wrong TRUDGE type computation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

> But i have a question about XC functionals, why always pure functionals are much more difficult to converge than hybrid functionals?

It's quite a typical situation, however for many systems,
pure functionals actually converges better.

The "exact" HF exchange has some unique properties,
which seriously distinguish it from GGA-type exchange.
This results in different convergence behavior of
pure HF, pure DFT, and hybrid functionals.


Message d4c0ee0abHW-6892-1392+00.htm, number 618, was posted on Thu Nov 13 '08 at 23:11:57
in reply to c3ce4a82XLp-6891-1158-00.htm

Re: PCGAMESS and RESP CHARGES

Alex Granovsky
gran@classic.chem.msu.su


Hi,

add $pdc constr=none $end to your input.

Regards,
Alex Granovsky



On Wed Nov 12 '08 7:26pm, Patrick wrote
---------------------------------------
>Hi PCGAMESS Forum,

>I would like to generate RESP charges for some small molecules using
>PCGAMESS as the generator of the ESP points that are the input to the RESP
>program. I've found instructions for doing this with GAMESS in the distribution
>of AMBER9. I've got the latest version of PCGAMESS and I've assumed that it supports
>this capability. Here are the relevant instructions for setting up the GAMESS input
>that I used:
> ****************************************************************************
> *******
> 1) Generate an appropiate GAMESS *.dat file which includes both the esp
> points and the atomic coordinates. The GAMESS input file which does
> this is shown here:

> $CONTRL SCFTYP=RHF EXETYP=RUN RUNTYP=OPTIMIZE COORD=UNIQUE $END
> $CONTRL MOLPLT=.TRUE. $END
> $STATPT NSTEP=100 $END
> $BASIS GBASIS=N31 NGAUSS=6 NDFUNC=1 $END
> $ELPOT IEPOT=1 WHERE=PDC OUTPUT=PUNCH $END
> $PDC PTSEL=CONNOLLY $END
> $GUESS GUESS=HUCKEL $END
> $DATA
> eg180.pdb
> C1
> C01 6.0 -0.6388649947 0.4052483921 0.0000000000
> ...
> ...
> H10 1.0 2.5277794177 0.0322897353 0.0000000000
> $END

>The runs completed successfully but did not produce the set of expected esp charges. Has anyone experienced similar issues? Is someone out there that has actually generated RESP charges using PCGAMESS to get esp points?

>I am grateful for any hint regarding this issue.

>Patrick
>


Message 93a30118ZML-6893-1376-00.htm, number 619, was posted on Fri Nov 14 '08 at 22:56:23
IRC computations with Firefly

Maurizio Ciofalo
omega@unipa.it


Dear all,
I noticed that Firefly is not able to keep trace of the OPTTOL keyword ($IRC group), though it is explicitly reported in the official PC GAMESS manual of June, 12, 1999 (the last shared with Firefly, I suppose). In other words, Firefly seems to ignore the value of OPTTOL provided by the user, at least within the GS2 pace which I have tested, as I can see in the output file.
Therefore it would be virtually impossibile at present to improve the accuracy of the convergence criteria for difficult cases, as Prof. Bernhard Schlegel suggests to do.
Is there anybody else who faced that problem?
Thank you very much for your attention.
Maurizio

Message d4c0ee0abHW-6894-47+00.htm, number 620, was posted on Sat Nov 15 '08 at 00:48:17
in reply to 93a30118ZML-6893-1376-00.htm

Re: IRC computations with Firefly

Alex Granovsky
gran@classic.chem.msu.su


Dear Maurizio,

The IRC code is designed so that it performs one or multiple
steps of the fixed step length.
Indeed, opttol is used by the GS2 code, however, the goal of the
IRC code is not to make gradient lower than opttol but rather to
perform next step along IRC. If gradient happens to fall below
opttol this is on of the possible criteria to stop IRC calculations.
However, this can only happen by chance. The most common criterium
to stop IRC calculation is thus the increase of the energy as
compared with the previosly found point on IRC.

I think the documentation is definitely not quite exact
concerning this point. However, if you set stride to some
very small value, you'll increase the probability to fulfill
the opttol-based criterium while walking along IRC; naturally
this would require much more compitations.

Regards,
Alex

On Fri Nov 14 '08 10:56pm, Maurizio Ciofalo wrote
-------------------------------------------------
>Dear all,
>I noticed that Firefly is not able to keep trace of the OPTTOL keyword ($IRC group), though it is explicitly reported in the official PC GAMESS manual of June, 12, 1999 (the last shared with Firefly, I suppose). In other words, Firefly seems to ignore the value of OPTTOL provided by the user, at least within the GS2 pace which I have tested, as I can see in the output file.
>Therefore it would be virtually impossibile at present to improve the accuracy of the convergence criteria for difficult cases, as Prof. Bernhard Schlegel suggests to do.
>Is there anybody else who faced that problem?
>Thank you very much for your attention.
>Maurizio
>


Message 4f0caf05ZML-6894-94+00.htm, number 621, was posted on Sat Nov 15 '08 at 01:33:45
in reply to d4c0ee0abHW-6894-47+00.htm

Re^2: IRC computations with Firefly

Maurizio Ciofalo
omega@unipa.it


On Sat Nov 15 '08 0:48am, Alex Granovsky wrote

>However, if you set stride to some
>very small value, you'll increase the probability to fulfill
>the opttol-based criterium while walking along IRC; naturally
>this would require much more compitations.

That is exactly the point which I would like to address. Sometimes one faces flat (i.e. small curvature) PES and therefore one needs to choose *larger* STRIDE values. At the same time, if one wants to avoid bad stops due to the flatness of the IRC, one should set the OPTTOL threshold to *smaller* values. Therefore IMHO the OPTTOL value should also be tunable by the Firefly users.

Regards



Message dcd10513NCe-6895-334-00.htm, number 622, was posted on Sun Nov 16 '08 at 05:36:17
Geometry optimization of S1 state by MCSCF

Fumihito Mohri
qzg00677@nifty.com


Dear PC-GAMESS users

I am studying MCSCF method. My final target is computation of
spin-orbit coupling constants for organic optical matetials.
Now, I am facing a problem with MCSCF job for HCHO molecule(this is an
excersize job).
Attached file is an input file. This HCHO molecule is not co-planar.
$VEC data was obtained by RHF/6-31+G(d) for this molecule. The number
of electrons is 16, thus HOMO is the 8th MO. The active space consists
of 7th, 8th, 9th and 10th MOs (i.e. CASSCF(4,4)). Symmetry is chosen
as C1. From this file, I want to obtain optimized geometry and energy
of S1 (the first excited singlet state). But, this input file brings
only S0 result (i.e. ground state result). Could you tell me what is
wrong with this file. I do not yet well understand the relationship
between NSTATE, IROOT and SZ.
I have one more question. In Gaussian, MO alternation is sometimes done
by Guess=Alter to make an appropriate active space. Does PC-Gamess have
the keyword corresponding to Guess=Alter?

Thank you in advance

This message contains the 30 kb attachment
[ HCHO-MCSCF-S1.inp ] HCHO-MCSCF-S1


Message 53d998acysT-6895-769-00.htm, number 623, was posted on Sun Nov 16 '08 at 12:49:00
Geometry restraints during optimisation

Miro Moman
edelmiro.moman@gmail.com


Hello,

I was wondering if there is a way to include restraints during a geometry optimisation in Firefly.

The context is the following: we have a substrate which is converted to a free radical by an enzyme. We have studied the reaction by EPR and we observed the shielding of two hydrogen atoms of the substrate. We have done some docking calculations and we believe we can explain this shielding because of the presence of a carboxylate residue in the vicinity of the hydrogens. We wanted to verify this by doing some DFT calculations and checking how the presence of the carboxylate affects the electron density distribution on the substrate.

I have carried out a geometry optimisation of the closed shell species (not the free radicals yet) at the HF/6-31G* level of theory and a single point calculation at the B3LYP/6++311G** level of theory, both with the substrate alone and with the substrate plus carboxylate complex. I am new to Firefly and I am an organic chemist without much experience in ab initio calculations.

The problem is that during the geometry optimisation the geometry of the complex diverges too much from that found in the docking calculation. The reason is that the position of the carboxylate is shifted to from a hydrogen bond with a nearby hydroxyl group of the substrate (this situation is not possible in the context of the active site of the enzyme).

In summary, I would like to know which options do exist to perform the geometry optimisation without significantly altering the relative orientation of the two molecules. Would it be fine just using lousier convergence criteria?

Best regards,

Miro


Message 53d998acysT-6895-769+00.htm, number 623, was edited on Sun Nov 16 '08 at 13:03:53
and replaces message 53d998acysT-6895-769-00.htm

Geometry restraints during optimisation

Miro Moman
edelmiro.moman@gmail.com


Hello,

I was wondering if there is a way to include restraints during a geometry optimisation in Firefly.

The context is the following: we have a substrate which is converted to a free radical by an enzyme. We have studied the reaction by EPR and we observed the shielding of two hydrogen atoms of the substrate. We have done some docking calculations and we believe we can explain this shielding because of the presence of a carboxylate residue in the vicinity of the hydrogens. We wanted to verify this by doing some DFT calculations and checking how the presence of the carboxylate affects the electron density distribution on the substrate.

I have carried out a geometry optimisation of the closed shell species (not the free radicals yet) at the HF/6-31G* level of theory and a single point calculation at the B3LYP/6++311G** level of theory, both with the substrate alone and with the substrate plus carboxylate complex. I am new to Firefly and I am an organic chemist without much experience in ab initio calculations.

The problem is that during the geometry optimisation the geometry of the complex diverges too much from that found in the docking calculation. The reason is that the position of the carboxylate is shifted to from a hydrogen bond with a nearby hydroxyl group of the substrate (this situation is not possible in the context of the active site of the enzyme).

In summary, I would like to know which options do exist to perform the geometry optimisation without significantly altering the relative orientation of the two molecules. Would it be fine just using looser convergence criteria?

Best regards,

Miro

[ This message was edited on Sun Nov 16 '08 at 1:03pm by the author ]


Message 4df668afbHW-6895-1195+00.htm, number 624, was posted on Sun Nov 16 '08 at 19:55:33
in reply to 4f0caf05ZML-6894-94+00.htm

Re^3: IRC computations with Firefly

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>That is exactly the point which I would like to address. Sometimes one faces flat (i.e. small curvature) PES and therefore one needs to choose *larger* STRIDE values. At the same time, if one wants to avoid bad stops due to the flatness of the IRC, one should set the OPTTOL threshold to *smaller* values. Therefore IMHO the OPTTOL value should also be tunable by the Firefly users.

The opttol variable is tunable. The point is that
(as I already explained) it usually has very little effect on IRC calculations.

On a separat note, are you sure you are really interested
in changing the opttol value? rather than just the values of
rtol/gtol parameters?

Regards,
Alex


Message 4df668afbHW-6895-1201+00.htm, number 625, was posted on Sun Nov 16 '08 at 20:01:09
in reply to 53d998acysT-6895-769+00.htm

Re: Geometry restraints during optimisation

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>I was wondering if there is a way to include restraints during a geometry optimisation in Firefly.

>In summary, I would like to know which options do exist to perform the geometry optimisation without significantly altering the relative orientation of the two molecules. Would it be fine just using looser convergence criteria?

Sure there are several ways to do this, see e.g. my recent reply to the similar question at:

http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-6892-1355+00.htm

Most of this information can be found in the manual,
with some additional tricks discussed on this forum.

Regards,
Alex Granovsky


Message 3ac0208d595-6895-1269-00.htm, number 626, was posted on Sun Nov 16 '08 at 21:09:10
TDDFT options for help

Yongle Li
yongleli@itcc.nju.edu.cn


Hi, I am a new user of PC-GAMESS/Firefly, when I tested the TDDFT options
recently I received errors, so I need some help for fixing this problem.

The computer I used is:
CPU: Intel-Xeon,
System: Red Hat Enterprise Linux AS release 4
Intel MPI 3.0, Intel MKL 9.0
The input file is:

$CONTRL SCFTYP=RHF RUNTYP=ENERGY DFTTYP=B3LYP CITYP=TDDFT $END
$TDDFT MULT=1 NSTATE=4 $END
$BASIS GBASIS=n31 ngauss=6 NDFUNC=1 $END
$DATA
H2O
CNV 2

O ?????8.0 ?0.0000000000 ?0.0000000000 ?0.7205815395
H ?????1.0 ?0.0000000000 ?0.7565140024 ?0.1397092302
$END

and when I subbed the job with:
mpirun < /dev/null -n 4 ./pcgamess -i test3.inp -o test3.out /localhome/yongle/PC-GAMESS

I received the error information:
========================================
TID 9329 caught signal 11, exiting.

Dump of registers follows


TID 9328 caught signal 11, exiting.

Dump of registers follows

eax :: 0x01de9d60, edx :: 0xf511d3c8

TID 9330 caught signal 11, exiting.

Dump of registers follows

ecx :: 0x0000003e, ebx :: 0xf5103f78
eax :: 0x1ff67478, esi :: 0xffffa89c, edi :: 0xf514bc40
edx :: 0xf511d3c8
ebp :: 0xffffa4b0, esp :: 0xffffa2b0
eip :: 0x10571717, ecx :: 0x00000002, eflags :: 0x00010286

ebx :: 0xf5103f78
cs :: 0x0023
ds :: 0x002b
es :: 0x002b
esi :: 0xffffa89c, ss :: 0x002b
fs :: 0x01cf
edi :: 0xf514bc40
gs :: 0x005b

Stack backtrace

ebp :: 0xffffa4b0, esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
esp :: 0xffffa2b0
esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
eip :: 0x10571717, eflags :: 0x00010286

cs :: 0x0023
ds :: 0x002b
es :: 0x002b
eax :: 0x1ff67478, ss :: 0x002b
fs :: 0x01cf
gs :: 0x005b

Stack backtrace

esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
edx :: 0xf511d3c8
esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
ecx :: 0x00000002, ebx :: 0xf5103f78
esi :: 0xffffa89c, edi :: 0xf514bc40
ebp :: 0xffffa4b0, esp :: 0xffffa2b0
eip :: 0x10571717, eflags :: 0x00010286

cs :: 0x0023
ds :: 0x002b
es :: 0x002b
ss :: 0x002b
fs :: 0x01cf
gs :: 0x005b

Stack backtrace

esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453

[yongle@cu02 PC-GAMESS]$ rank 2 in job 1 localhost_49714 caused collective abort of all ranks
exit status of rank 2: killed by signal 9
=====================================================

And in output file:

--------------------------
R-B3LYP SCF CALCULATION
--------------------------

NUCLEAR ENERGY = 9.2267391005
MAXIT = 30 NPUNCH= 2
EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
DENSITY CONV= 1.00E-05
MEMORY REQUIRED FOR SCF STEP= 85597 WORDS.

XC FUNCTIONAL: SLATER + BECKE 88 + HF EXCHANGE, LYP 88 + VWN 1 RPA CORRELATION


ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR
* * * INITIATING DIIS PROCEDURE * * *
1 0 0 -97.011922222 -97.011922222 57.438108747 16.002497901


Message 53d998acysT-6895-1279+00.htm, number 627, was posted on Sun Nov 16 '08 at 21:19:41
in reply to 4df668afbHW-6895-1201+00.htm

Re^2: Geometry restraints during optimisation

Miro Moman
edelmiro.moman@gmail.com


>Most of this information can be found in the manual,
>with some additional tricks discussed on this forum.

Thanks Alex, I have already started a calculation which should
"freeze" the coordinates of 3 atoms (2 carbons in molecule A and 1
oxygen in molecule B). However, this man not be an optimal solution in
this case because those coordinates come from a docking calculation
using a united FF and I think the two molecules are a bit too close to
each other.

>Regards,
>Alex Granovsky
>


Message 4df668afbHW-6895-1332+00.htm, number 628, was posted on Sun Nov 16 '08 at 22:12:35
in reply to 53d998acysT-6895-1279+00.htm

Re^3: Geometry restraints during optimisation

Alex Granovsky
gran@classic.chem.msu.su


>Thanks Alex, I have already started a calculation which should
>"freeze" the coordinates of 3 atoms (2 carbons in molecule A and 1
>oxygen in molecule B). However, this man not be an optimal solution in
>this case because those coordinates come from a docking calculation
>using a united FF and I think the two molecules are a bit too close to
>each other.

It is usually better to define z-matrix, either manually or automatically,
and then fix internal coordinates. This is much more flexible approach.

Regards,
Alex


Message 4df668afbHW-6895-1353+00.htm, number 629, was posted on Sun Nov 16 '08 at 22:33:05
in reply to 3ac0208d595-6895-1269-00.htm

Re: TDDFT options for help

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your mistake is that you did not assign working directories
for different instances of the parallel PC GAMESS/Firefly process.

The simplest way to do this is to use -t command line option.

As it is not documented very well in the manuals, I post here the details on how it works.


Syntax:

pcgamess [other options] -t < the tmp dirs prefix >

Examples:

Case 1 (formally correct but deadly):

 -t /tmp 

Each instance will try to create /tmp. directory for its scratch files. If you are not running as root, they will fail resulting in different runtime errors depending on the error
code returned by OS and the filesystem type.

Case 2 (correct use, note the difference with case 1!):

 -t /tmp/ 

Each instance will try to create /tmp/pcg. directory for its scratch files.

Case 3 (correct use):

 -t /tmp/test 

Each instance will try to create /tmp/test. directory for its scratch files.

Case 4 (correct use):

 -t /tmp/test/ 

Each instance will try to create /tmp/test/pcg. directory for its scratch files.


In addition, I would recommend you to specify -r and -p command line
switches for most of the runs. Finally, the proper -ex option is
important for better performance. And remember that all options which
take path/filename as their argument (including -i and -o options)
expect fully qualified filename or path. This is especially important
because the PC GAMESS/Firefly' working directory is not usually the
directory it was launched from!


-Alex Granovsky









On Sun Nov 16 '08 9:09pm, Yongle Li wrote
-----------------------------------------
>Hi, I am a new user of PC-GAMESS/Firefly, when I tested the TDDFT options
> recently I received errors, so I need some help for fixing this problem.

>The computer I used is:
>CPU: Intel-Xeon,
>System: Red Hat Enterprise Linux AS release 4
>Intel MPI 3.0, Intel MKL 9.0
>The input file is:

> $CONTRL SCFTYP=RHF RUNTYP=ENERGY DFTTYP=B3LYP CITYP=TDDFT $END
> $TDDFT MULT=1 NSTATE=4 $END
> $BASIS GBASIS=n31 ngauss=6 NDFUNC=1 $END
> $DATA
>H2O
>CNV 2

> O ?????8.0 ?0.0000000000 ?0.0000000000 ?0.7205815395
> H ?????1.0 ?0.0000000000 ?0.7565140024 ?0.1397092302
> $END

>and when I subbed the job with:
>mpirun < /dev/null -n 4 ./pcgamess -i test3.inp -o test3.out /localhome/yongle/PC-GAMESS

>I received the error information:
>========================================
>TID 9329 caught signal 11, exiting.

>Dump of registers follows
>
>
>TID 9328 caught signal 11, exiting.

>Dump of registers follows

>eax :: 0x01de9d60, edx :: 0xf511d3c8

>TID 9330 caught signal 11, exiting.

>Dump of registers follows

>ecx :: 0x0000003e, ebx :: 0xf5103f78
>eax :: 0x1ff67478, esi :: 0xffffa89c, edi :: 0xf514bc40
>edx :: 0xf511d3c8
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, ecx :: 0x00000002, eflags :: 0x00010286

>ebx :: 0xf5103f78
>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>esi :: 0xffffa89c, ss :: 0x002b
>fs :: 0x01cf
>edi :: 0xf514bc40
>gs :: 0x005b

>Stack backtrace

>ebp :: 0xffffa4b0, esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa2b0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>eax :: 0x1ff67478, ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>edx :: 0xf511d3c8
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>ecx :: 0x00000002, ebx :: 0xf5103f78
>esi :: 0xffffa89c, edi :: 0xf514bc40
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453

>[yongle@cu02 PC-GAMESS]$ rank 2 in job 1 localhost_49714 caused collective abort of all ranks
> exit status of rank 2: killed by signal 9
>=====================================================

>And in output file:

> --------------------------
> R-B3LYP SCF CALCULATION
> --------------------------

> NUCLEAR ENERGY = 9.2267391005
> MAXIT = 30 NPUNCH= 2
> EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
> DENSITY CONV= 1.00E-05
> MEMORY REQUIRED FOR SCF STEP= 85597 WORDS.

> XC FUNCTIONAL: SLATER + BECKE 88 + HF EXCHANGE, LYP 88 + VWN 1 RPA CORRELATION
>
>
> ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR
> * * * INITIATING DIIS PROCEDURE * * *
> 1 0 0 -97.011922222 -97.011922222 57.438108747 16.002497901

>


Message 4df668afbHW-6895-1354+00.htm, number 629, was edited on Sun Nov 16 '08 at 22:34:29
and replaces message 4df668afbHW-6895-1353+00.htm

Re: TDDFT options for help

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your mistake is that you did not assign working directories
for different instances of the parallel PC GAMESS/Firefly process.

The simplest way to do this is to use -t command line option.

As it is not documented very well in the manuals, I post here the details on how it works.


Syntax:

pcgamess [other options] -t < the tmp dirs prefix >

Examples:

Case 1 (formally correct but deadly):

 -t /tmp 

Each instance will try to create /tmp. directory for
its scratch files. If you are not running as root, they will fail
resulting in different runtime errors depending on the error
code returned by OS and the filesystem type.

Case 2 (correct use, note the difference with case 1!):

 -t /tmp/ 

Each instance will try to create /tmp/pcg.
directory for its scratch files.

Case 3 (correct use):

 -t /tmp/test 

Each instance will try to create /tmp/test.
directory for its scratch files.

Case 4 (correct use):

 -t /tmp/test/ 

Each instance will try to create /tmp/test/pcg.
directory for its scratch files.

etc...


In addition, I would recommend you to specify -r and -p command line
switches for most of the runs. Finally, the proper -ex option is
important for better performance. And remember that all options which
take path/filename as their argument (including -i and -o options)
expect fully qualified filename or path. This is especially important
because the PC GAMESS/Firefly' working directory is not usually the
directory it was launched from!


-Alex Granovsky









On Sun Nov 16 '08 9:09pm, Yongle Li wrote
-----------------------------------------
>Hi, I am a new user of PC-GAMESS/Firefly, when I tested the TDDFT options
> recently I received errors, so I need some help for fixing this problem.

>The computer I used is:
>CPU: Intel-Xeon,
>System: Red Hat Enterprise Linux AS release 4
>Intel MPI 3.0, Intel MKL 9.0
>The input file is:

> $CONTRL SCFTYP=RHF RUNTYP=ENERGY DFTTYP=B3LYP CITYP=TDDFT $END
> $TDDFT MULT=1 NSTATE=4 $END
> $BASIS GBASIS=n31 ngauss=6 NDFUNC=1 $END
> $DATA
>H2O
>CNV 2

> O ?????8.0 ?0.0000000000 ?0.0000000000 ?0.7205815395
> H ?????1.0 ?0.0000000000 ?0.7565140024 ?0.1397092302
> $END

>and when I subbed the job with:
>mpirun < /dev/null -n 4 ./pcgamess -i test3.inp -o test3.out /localhome/yongle/PC-GAMESS

>I received the error information:
>========================================
>TID 9329 caught signal 11, exiting.

>Dump of registers follows
>
>
>TID 9328 caught signal 11, exiting.

>Dump of registers follows

>eax :: 0x01de9d60, edx :: 0xf511d3c8

>TID 9330 caught signal 11, exiting.

>Dump of registers follows

>ecx :: 0x0000003e, ebx :: 0xf5103f78
>eax :: 0x1ff67478, esi :: 0xffffa89c, edi :: 0xf514bc40
>edx :: 0xf511d3c8
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, ecx :: 0x00000002, eflags :: 0x00010286

>ebx :: 0xf5103f78
>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>esi :: 0xffffa89c, ss :: 0x002b
>fs :: 0x01cf
>edi :: 0xf514bc40
>gs :: 0x005b

>Stack backtrace

>ebp :: 0xffffa4b0, esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa2b0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>eax :: 0x1ff67478, ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>edx :: 0xf511d3c8
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>ecx :: 0x00000002, ebx :: 0xf5103f78
>esi :: 0xffffa89c, edi :: 0xf514bc40
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453

>[yongle@cu02 PC-GAMESS]$ rank 2 in job 1 localhost_49714 caused collective abort of all ranks
> exit status of rank 2: killed by signal 9
>=====================================================

>And in output file:

> --------------------------
> R-B3LYP SCF CALCULATION
> --------------------------

> NUCLEAR ENERGY = 9.2267391005
> MAXIT = 30 NPUNCH= 2
> EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
> DENSITY CONV= 1.00E-05
> MEMORY REQUIRED FOR SCF STEP= 85597 WORDS.

> XC FUNCTIONAL: SLATER + BECKE 88 + HF EXCHANGE, LYP 88 + VWN 1 RPA CORRELATION
>
>
> ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR
> * * * INITIATING DIIS PROCEDURE * * *
> 1 0 0 -97.011922222 -97.011922222 57.438108747 16.002497901

>

[ This message was edited on Sun Nov 16 '08 at 10:34pm by the author ]


Message 4df668afbHW-6895-1357+00.htm, number 629, was edited on Sun Nov 16 '08 at 22:37:14
and replaces message 4df668afbHW-6895-1354+00.htm

Re: TDDFT options for help

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your mistake is that you did not assign working directories
for different instances of the parallel PC GAMESS/Firefly process.

The simplest way to do this is to use -t command line option.

As it is not documented very well in the manuals, I post here the details on how it works.


Syntax:

pcgamess [other options] -t < the tmp dirs prefix >

Examples:

Case 1 (formally correct but deadly):

 -t /tmp 

Each instance will try to create /tmp.process_rank directory for
its scratch files. If you are not running as root, they will fail
resulting in different runtime errors depending on the error
code returned by OS and the filesystem type.

Case 2 (correct use, note the difference with case 1!):

 -t /tmp/ 

Each instance will try to create /tmp/pcg.process_rank
directory for its scratch files, e.g., /tmp/pcg.0. /tmp/pcg.1, etc...


Case 3 (correct use):

 -t /tmp/test 

Each instance will try to create /tmp/test.process_rank
directory for its scratch files.

Case 4 (correct use):

 -t /tmp/test/ 

Each instance will try to create /tmp/test/pcg.process_rank
directory for its scratch files.

etc..., etc...


In addition, I would recommend you to specify -r and -p command line
switches for most of the runs. Finally, the proper -ex option is
important for better performance. And remember that all options which
take path/filename as their argument (including -i and -o options)
expect fully qualified filename or path. This is especially important
because the PC GAMESS/Firefly' working directory is usually not the
directory it was launched from!

-Alex Granovsky









On Sun Nov 16 '08 9:09pm, Yongle Li wrote
-----------------------------------------
>Hi, I am a new user of PC-GAMESS/Firefly, when I tested the TDDFT options
> recently I received errors, so I need some help for fixing this problem.

>The computer I used is:
>CPU: Intel-Xeon,
>System: Red Hat Enterprise Linux AS release 4
>Intel MPI 3.0, Intel MKL 9.0
>The input file is:

> $CONTRL SCFTYP=RHF RUNTYP=ENERGY DFTTYP=B3LYP CITYP=TDDFT $END
> $TDDFT MULT=1 NSTATE=4 $END
> $BASIS GBASIS=n31 ngauss=6 NDFUNC=1 $END
> $DATA
>H2O
>CNV 2

> O ?????8.0 ?0.0000000000 ?0.0000000000 ?0.7205815395
> H ?????1.0 ?0.0000000000 ?0.7565140024 ?0.1397092302
> $END

>and when I subbed the job with:
>mpirun < /dev/null -n 4 ./pcgamess -i test3.inp -o test3.out /localhome/yongle/PC-GAMESS

>I received the error information:
>========================================
>TID 9329 caught signal 11, exiting.

>Dump of registers follows
>
>
>TID 9328 caught signal 11, exiting.

>Dump of registers follows

>eax :: 0x01de9d60, edx :: 0xf511d3c8

>TID 9330 caught signal 11, exiting.

>Dump of registers follows

>ecx :: 0x0000003e, ebx :: 0xf5103f78
>eax :: 0x1ff67478, esi :: 0xffffa89c, edi :: 0xf514bc40
>edx :: 0xf511d3c8
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, ecx :: 0x00000002, eflags :: 0x00010286

>ebx :: 0xf5103f78
>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>esi :: 0xffffa89c, ss :: 0x002b
>fs :: 0x01cf
>edi :: 0xf514bc40
>gs :: 0x005b

>Stack backtrace

>ebp :: 0xffffa4b0, esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa2b0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>eax :: 0x1ff67478, ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>edx :: 0xf511d3c8
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>ecx :: 0x00000002, ebx :: 0xf5103f78
>esi :: 0xffffa89c, edi :: 0xf514bc40
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453

>[yongle@cu02 PC-GAMESS]$ rank 2 in job 1 localhost_49714 caused collective abort of all ranks
> exit status of rank 2: killed by signal 9
>=====================================================

>And in output file:

> --------------------------
> R-B3LYP SCF CALCULATION
> --------------------------

> NUCLEAR ENERGY = 9.2267391005
> MAXIT = 30 NPUNCH= 2
> EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
> DENSITY CONV= 1.00E-05
> MEMORY REQUIRED FOR SCF STEP= 85597 WORDS.

> XC FUNCTIONAL: SLATER + BECKE 88 + HF EXCHANGE, LYP 88 + VWN 1 RPA CORRELATION
>
>
> ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR
> * * * INITIATING DIIS PROCEDURE * * *
> 1 0 0 -97.011922222 -97.011922222 57.438108747 16.002497901

>

[ This message was edited on Sun Nov 16 '08 at 10:37pm by the author ]


Message 4df668afbHW-6895-1360+00.htm, number 629, was edited on Sun Nov 16 '08 at 22:39:49
and replaces message 4df668afbHW-6895-1357+00.htm

Re: TDDFT options for help

Alex Granovsky
gran@classic.chem.msu.su


Hi,

your mistake is that you did not assign working directories
for different instances of the parallel PC GAMESS/Firefly process.

The simplest way to do this is to use -t command line option.

As it is not documented very well in the manuals, I post here the details on how it works.


Syntax:

pcgamess [other options] -t < the tmp dirs prefix >

Examples:

Case 1 (formally correct but deadly):

 -t /tmp 

Each instance will try to create /tmp.process_rank directory for
its scratch files, i.e., /tmp.0, /tmp.1, etc... If you are not
running as root, they will fail resulting in different runtime
errors depending on the error code returned by OS and the
filesystem type.

Case 2 (correct use, note the difference with Case 1!):

 -t /tmp/ 

Each instance will try to create /tmp/pcg.process_rank
directory for its scratch files, i.e., /tmp/pcg.0. /tmp/pcg.1, etc...


Case 3 (correct use):

 -t /tmp/test 

Each instance will try to create /tmp/test.process_rank
directory for its scratch files.

Case 4 (correct use):

 -t /tmp/test/ 

Each instance will try to create /tmp/test/pcg.process_rank
directory for its scratch files.

etc..., etc...


In addition, I would recommend you to specify -r and -p command line
switches for most of the runs. Finally, the proper -ex option is
important for better performance. And remember that all options which
take path/filename as their argument (including -i and -o options)
expect fully qualified filename or path. This is especially important
because the PC GAMESS/Firefly' working directory is usually not the
directory it was launched from!

-Alex Granovsky









On Sun Nov 16 '08 9:09pm, Yongle Li wrote
-----------------------------------------
>Hi, I am a new user of PC-GAMESS/Firefly, when I tested the TDDFT options
> recently I received errors, so I need some help for fixing this problem.

>The computer I used is:
>CPU: Intel-Xeon,
>System: Red Hat Enterprise Linux AS release 4
>Intel MPI 3.0, Intel MKL 9.0
>The input file is:

> $CONTRL SCFTYP=RHF RUNTYP=ENERGY DFTTYP=B3LYP CITYP=TDDFT $END
> $TDDFT MULT=1 NSTATE=4 $END
> $BASIS GBASIS=n31 ngauss=6 NDFUNC=1 $END
> $DATA
>H2O
>CNV 2

> O ?????8.0 ?0.0000000000 ?0.0000000000 ?0.7205815395
> H ?????1.0 ?0.0000000000 ?0.7565140024 ?0.1397092302
> $END

>and when I subbed the job with:
>mpirun < /dev/null -n 4 ./pcgamess -i test3.inp -o test3.out /localhome/yongle/PC-GAMESS

>I received the error information:
>========================================
>TID 9329 caught signal 11, exiting.

>Dump of registers follows
>
>
>TID 9328 caught signal 11, exiting.

>Dump of registers follows

>eax :: 0x01de9d60, edx :: 0xf511d3c8

>TID 9330 caught signal 11, exiting.

>Dump of registers follows

>ecx :: 0x0000003e, ebx :: 0xf5103f78
>eax :: 0x1ff67478, esi :: 0xffffa89c, edi :: 0xf514bc40
>edx :: 0xf511d3c8
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, ecx :: 0x00000002, eflags :: 0x00010286

>ebx :: 0xf5103f78
>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>esi :: 0xffffa89c, ss :: 0x002b
>fs :: 0x01cf
>edi :: 0xf514bc40
>gs :: 0x005b

>Stack backtrace

>ebp :: 0xffffa4b0, esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa2b0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>eax :: 0x1ff67478, ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>edx :: 0xf511d3c8
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453
>ecx :: 0x00000002, ebx :: 0xf5103f78
>esi :: 0xffffa89c, edi :: 0xf514bc40
>ebp :: 0xffffa4b0, esp :: 0xffffa2b0
>eip :: 0x10571717, eflags :: 0x00010286

>cs :: 0x0023
>ds :: 0x002b
>es :: 0x002b
>ss :: 0x002b
>fs :: 0x01cf
>gs :: 0x005b

>Stack backtrace

>esp :: 0xffffa4b0, ebp :: 0xffffa798, eip :: 0xf50f07e0
>esp :: 0xffffa798, ebp :: 0x0b3a2b69, eip :: 0xc01e5453

>[yongle@cu02 PC-GAMESS]$ rank 2 in job 1 localhost_49714 caused collective abort of all ranks
> exit status of rank 2: killed by signal 9
>=====================================================

>And in output file:

> --------------------------
> R-B3LYP SCF CALCULATION
> --------------------------

> NUCLEAR ENERGY = 9.2267391005
> MAXIT = 30 NPUNCH= 2
> EXTRAP=T DAMP=T SHIFT=T RSTRCT=F DIIS=T DEM=F SOSCF=F
> DENSITY CONV= 1.00E-05
> MEMORY REQUIRED FOR SCF STEP= 85597 WORDS.

> XC FUNCTIONAL: SLATER + BECKE 88 + HF EXCHANGE, LYP 88 + VWN 1 RPA CORRELATION
>
>
> ITER EX DEM TOTAL ENERGY E CHANGE DENSITY CHANGE DIIS ERROR
> * * * INITIATING DIIS PROCEDURE * * *
> 1 0 0 -97.011922222 -97.011922222 57.438108747 16.002497901

>

[ This message was edited on Sun Nov 16 '08 at 10:39pm by the author ]


Message 4df668afbHW-6895-1372+00.htm, number 630, was posted on Sun Nov 16 '08 at 22:53:03
in reply to dcd10513NCe-6895-334-00.htm

Re: Geometry optimization of S1 state by MCSCF

Alex Granovsky
gran@classic.chem.msu.su


Hello,

I've attached the samples of CASSCF (2,2) geometry optimization
of S0 and S1 for your system using both state-specific and state-
averaged MCSCF.

Note that for S1 with state-spcific CASSCF, optimization failed
as CASSCF did not converge - this is quite a common situation
for excited states. In this particular case, the reason of
failure is the quasi-degeneracy of S0 and S1 at some geometries.

I'd then recommend you to use SA-CASSCF, which is much more stable.

The proper selection of active space is very important.
CAS (2,2) is natural for your problem. You can use larger spaces
but be careful to select proper orbitals!

And sure, you can use $guess norder=1 iorder(1)=... to reorder orbitals.

Regards,
Alex Granovsky



On Sun Nov 16 '08 5:36am, Fumihito Mohri wrote
----------------------------------------------
>Dear PC-GAMESS users

>I am studying MCSCF method. My final target is computation of
>spin-orbit coupling constants for organic optical matetials.
>Now, I am facing a problem with MCSCF job for HCHO molecule(this is an
>excersize job).
> Attached file is an input file. This HCHO molecule is not co-planar.
>$VEC data was obtained by RHF/6-31+G(d) for this molecule. The number
> of electrons is 16, thus HOMO is the 8th MO. The active space consists
> of 7th, 8th, 9th and 10th MOs (i.e. CASSCF(4,4)). Symmetry is chosen
>as C1. From this file, I want to obtain optimized geometry and energy
>of S1 (the first excited singlet state). But, this input file brings
> only S0 result (i.e. ground state result). Could you tell me what is
> wrong with this file. I do not yet well understand the relationship
> between NSTATE, IROOT and SZ.
> I have one more question. In Gaussian, MO alternation is sometimes done
> by Guess=Alter to make an appropriate active space. Does PC-Gamess have
>the keyword corresponding to Guess=Alter?

>Thank you in advance
>

This message contains the 119 kb attachment
[ ch2o_s0_s1.rar ] SS and SA MCSCF geometry optimization samples for CH2O


Message 4df668afbHW-6895-1374+00.htm, number 630, was edited on Sun Nov 16 '08 at 22:53:50
and replaces message 4df668afbHW-6895-1372+00.htm

Re: Geometry optimization of S1 state by MCSCF

Alex Granovsky
gran@classic.chem.msu.su


Hello,

I've attached the samples of CASSCF (2,2) geometry optimization
of S0 and S1 states for your system using both state-specific and
state-averaged MCSCF.

Note that for S1 with state-spcific CASSCF, optimization failed
as CASSCF did not converge - this is quite a common situation
for excited states. In this particular case, the reason of
failure is the quasi-degeneracy of S0 and S1 at some geometries.

I'd then recommend you to use SA-CASSCF, which is much more stable.

The proper selection of active space is very important.
CAS (2,2) is natural for your problem. You can use larger spaces
but be careful to select proper orbitals!

And sure, you can use $guess norder=1 iorder(1)=... to reorder orbitals.

Regards,
Alex Granovsky



On Sun Nov 16 '08 5:36am, Fumihito Mohri wrote
----------------------------------------------
>Dear PC-GAMESS users

>I am studying MCSCF method. My final target is computation of
>spin-orbit coupling constants for organic optical matetials.
>Now, I am facing a problem with MCSCF job for HCHO molecule(this is an
>excersize job).
> Attached file is an input file. This HCHO molecule is not co-planar.
>$VEC data was obtained by RHF/6-31+G(d) for this molecule. The number
> of electrons is 16, thus HOMO is the 8th MO. The active space consists
> of 7th, 8th, 9th and 10th MOs (i.e. CASSCF(4,4)). Symmetry is chosen
>as C1. From this file, I want to obtain optimized geometry and energy
>of S1 (the first excited singlet state). But, this input file brings
> only S0 result (i.e. ground state result). Could you tell me what is
> wrong with this file. I do not yet well understand the relationship
> between NSTATE, IROOT and SZ.
> I have one more question. In Gaussian, MO alternation is sometimes done
> by Guess=Alter to make an appropriate active space. Does PC-Gamess have
>the keyword corresponding to Guess=Alter?

>Thank you in advance
>

This message contains the 119 kb attachment
[ ch2o_s0_s1.rar ] SS and SA MCSCF geometry optimization samples for CH2O

[ This message was edited on Sun Nov 16 '08 at 10:53pm by the author ]


Message 86605bbdysT-6896-745+00.htm, number 631, was posted on Mon Nov 17 '08 at 12:24:40
in reply to 4df668afbHW-6895-1332+00.htm

Re^4: Geometry restraints during optimisation

Miro Moman
edelmiro.moman@gmail.com


>It is usually better to define z-matrix, either manually or automatically,
>and then fix internal coordinates. This is much more flexible approach.

Sure. In this case I have two molecules, can one define some kind of intermolecular pseudo-bonds and constrain those?


Message dcd10513NCe-6896-1021+00.htm, number 632, was posted on Mon Nov 17 '08 at 17:03:16
in reply to 4df668afbHW-6895-1374+00.htm

Re^2: Geometry optimization of S1 state by MCSCF

Fumihito Mohri
qzg00677@nifty.com


Dear Prof. Granovsky,
Thank very much for your time and for example files. I have known that the state-
averaged MCSCF is used to avoid 'root-fliping'. Any way, I have some questions for
the example file of opt.s1.sa.inp (here,I confirm that absence of MULT in $CONTRL
means singlet-state computation).

>! Need extra states in CI as the second one is triplet
> $DET NCORE=7 NACT=2 NELS=2 NSTATE=4 GROUP=C1 ITERMX=200
>PURES=.T. $END
I understand that PURES=.T. means that only singlet states are included in the
state-averaging. Is this understanding correct? The sentence of 'the second one is
triplet' makes me confused. Why the triplet state appears in this case? Moreover,
how did you determine the number of states of four, that is, how did you determine
NSTATE=4 before doing this run?

! State-averaged MCSCF
$DET WSTATE(1)=1,1 $END
! istate=2 means we need properties and gradiends of second state (S1)
$mcscf CISTEP=aldet maxit=200 ISTATE=2 $end
The expression of 'second state (S1)' also makes me confused, because this seems to be
inconsistent with the above sentence 'the second one is triplet'. I imagine that the
triplet state is removed and thus the S1 state is 2nd lower state among all singlet
states. Is this correct?

In opt.s1.sa.out
STATE 2 ENERGY= -113.6672881517 S= 1.00 SZ= 0.00 SPACE SYM=A
ALPH|BETA| COEFFICIENT
----|----|------------
01 | 10 | -0.7071068
10 | 01 | 0.7071068
10 | 10 | 0.0000000

I think that the above combinations of ALPHA and BETA should be of singlet states.
But S= 1.00 is output. Why? (I think that this means 2S+1=3)

Best, regards,
Fumihito Mohri
Kaneka Corportaion, Japan


Message dcd10513NCe-6896-1065+00.htm, number 633, was posted on Mon Nov 17 '08 at 17:45:33
in reply to 53d998acysT-6895-769+00.htm

Re: Geometry restraints during optimisation

Fumihito Mohri
qzg00677@nifty.com


Dear Miro Moman

>The reason is that the position of the carboxylate is shifted to from a
>hydrogen bond with a nearby hydroxyl group of the substrate
Is it possible to make manually the hydrogen bond between the carboxylate
and the hydroxyl group with a reasonable O---O distance (around 2.7-2.8 Angs)?
Any way, generally speaking, in such a system, the geometry optimisation
usually brings ediverged structure', as you experienced. The solution of
this problem may be due to your purpose (i.e. what you want with this system).
If there are no objections, please show us the coordinates data of your system,
and you will receive more specific advice.

Best regards,
Fumihito Mohri


Message 4405ef556g1-6896-1074-00.htm, number 634, was posted on Mon Nov 17 '08 at 17:54:19
B3YLP5 Job with Dunning Basis Set

Thomas Patko
tpatko@gmail.com


I am trying to run some SPE jobs for my systems using the level theory provided by another paper as they had reasonable results. I need the Dunning basis sets as I have iodide as one of my atoms and the typical 6-31G* basis set will no support this.

The popular three-parameter hybrid functional of Becke (B3LYP) [22] was employed in this study in combination with the DGauss double-zeta valence polarization (DGDZVP) basis- set [23,24].

This job in this paper was run on Gaussian 03. I believe that the equivalent method in PCG would be B3YLP5. I am unsure how to get the same double zeta dunning basis set with the same polarization?

Thanks,

Thomas


Message 470d39a48Nm-6896-1076+00.htm, number 635, was posted on Mon Nov 17 '08 at 17:56:23
in reply to c0ace279U2I-6889-781-00.htm

Re: Firefly (PCG71C) compatibility with Tinker 4.2

Jim Kress
jimkress_35@kressworks.com


On Mon Nov 10 '08 1:00pm, Veinardi Suendo wrote
-----------------------------------------------
>Dear Sirs,
>How to do QM/MM calculation with PCGAMESS/Tinker. Is there any recent manual or guideline? Can I use Firefly (PCG71C) with Tinker 4.2? I have found an old manual for PCG65 that works only with Tinker 3.7(http://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htmhttp://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htmhttp://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htmhttp://classic.chem.msu.su/gran/gamess/forum/attach/QMMM%20input%20documentation%2001Aug07.htm). Or is there any other MM package for doing QM/MM calculation with Firefly.

>Thank you very much in advance,
>
>
>Yours Sincerely,
>
>
>Veinardi Suendo


The current version of Firefly does not have any QM/MM capability (other than effective fragment potentials - see the Firefly documentation for more details). The only version of PC GAMESS that has an integrated QM/MM capability is v6.4 and the MM part of that is Tinker 3.7. The manual to which you refer is the correct manual for the v6.4 QM/MM version of PC GAMESS.

We anticipate integrating MM capabilities into Firefly. Unfortunately, a schedule of when that will be completed has not been developed. It (the integration, not the schedule development) will certainly be done within the next 6 months or so.

The challenge is providing an integration that can keep up with the rapid pace of Alex's additions to/ improvements in Firefly ... :>}

I hope this helps.

Jim Kress


Message 4df668afbHW-6896-1285+00.htm, number 636, was posted on Mon Nov 17 '08 at 21:24:37
in reply to 4405ef556g1-6896-1074-00.htm

Re: B3YLP5 Job with Dunning Basis Set

Alex Granovsky
gran@classic.chem.msu.su


Hi Thomas,

look at:

http://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtml

I think you should use B3LYP1, not B3LYP5

Regards,
Alex


On Mon Nov 17 '08 5:54pm, Thomas Patko wrote
--------------------------------------------
>I am trying to run some SPE jobs for my systems using the level theory provided by another paper as they had reasonable results. I need the Dunning basis sets as I have iodide as one of my atoms and the typical 6-31G* basis set will no support this.

>The popular three-parameter hybrid functional of Becke (B3LYP) [22] was employed in this study in combination with the DGauss double-zeta valence polarization (DGDZVP) basis- set [23,24].

>This job in this paper was run on Gaussian 03. I believe that the equivalent method in PCG would be B3YLP5. I am unsure how to get the same double zeta dunning basis set with the same polarization?

>Thanks,

>Thomas


Message 4405ef556g1-6897-402+00.htm, number 637, was posted on Tue Nov 18 '08 at 06:42:18
in reply to 4df668afbHW-6896-1285+00.htm

Re^2: B3YLP5 Job with Dunning Basis Set

Thomas Patko
tpatko@gmail.com


Hello Alex:

Using MacMolPLT Input Builder to construct the basis set for a straight forward correlation consistent Dunning basis set like cc-pVDZ I used:

$BASIS GBASIS=CCD NDFUNC=1 $END

and added the ISPHER=1 option (as I believe would be required to use such a basis set). The ISPHER=1 option does not show as valid for PCG.

Even after removing the ISPHER=1, this always results in an error than an illegal basis function was requested. Are the Dunning correlation consistent basis sets implemented for PCG? If yes, can they be specified the same as they are in GAMESS-US jobs (for example how MacMolPLT would struct these basis sets)?

Thanks,

Thomas

On Mon Nov 17 '08 9:24pm, Alex Granovsky wrote
----------------------------------------------
>Hi Thomas,

>look at:

>http://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtmlhttp://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtml

>I think you should use B3LYP1, not B3LYP5

>Regards,
>Alex
>
>
>On Mon Nov 17 '08 5:54pm, Thomas Patko wrote
>--------------------------------------------
>>I am trying to run some SPE jobs for my systems using the level theory provided by another paper as they had reasonable results. I need the Dunning basis sets as I have iodide as one of my atoms and the typical 6-31G* basis set will no support this.

>>The popular three-parameter hybrid functional of Becke (B3LYP) [22] was employed in this study in combination with the DGauss double-zeta valence polarization (DGDZVP) basis- set [23,24].

>>This job in this paper was run on Gaussian 03. I believe that the equivalent method in PCG would be B3YLP5. I am unsure how to get the same double zeta dunning basis set with the same polarization?

>>Thanks,

>>Thomas


Message 4df668afbHW-6897-963+00.htm, number 638, was posted on Tue Nov 18 '08 at 16:03:38
in reply to 4405ef556g1-6897-402+00.htm

Re^3: B3YLP5 Job with Dunning Basis Set

Alex Granovsky
gran@classic.chem.msu.su


Hello Thomas,

I hope the following comments would be helpful:

1. The DGDZVP basis is NOT the member of Dunning's cc-type basis set
family. Rather, it's the basis set used by DGauss and is designed for
DFT calculations. It is availabe at:

http://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtml

but unfortunately in DGauss format. This format is documented
inside the documnts containing basis sets themselves. Note you
do not need to worry on auxiliary (fitting) basis sets at all.

However, you can write some script to format it as PC GAMESS/Firefly
external basis set file. See the cc-type basis sets available in the
"Downloads" section of our web server for samples of the external
basis set files. Please share it with us if you really create GAMESS
format external basis set library for DGDZVP.

2. The PC GAMESS does not store Dunning's cc-type basis
sets internally, however, they can be used using

 $basis extfil=.t. gbasis=name $end

directive. You may need to use -b command line option to set the
path to and the name of the basis set library file. There are
multiple files containng definitions of popular external basis
sets available in the "Downloads" section.

3. The GAMESS' ispher option is not supported, however,
the PC GAMESS-specific d5 option and related $d5 group
can be used instead. Together, they allow finer control on removal of
spherical contaminants. This option is documented in the "Manuals"
section:


http://classic.chem.msu.su/gran/gamess/spherical.html

4. Finally, if you use external basis sets, any of NDFUNC, etc... are ignored.

Regards,
Alex




Message 4df668afbHW-6897-965+00.htm, number 638, was edited on Tue Nov 18 '08 at 16:05:24
and replaces message 4df668afbHW-6897-963+00.htm

Re^3: B3YLP5 Job with Dunning Basis Set

Alex Granovsky
gran@classic.chem.msu.su


Hello Thomas,

I hope the following comments would be helpful:

1. The DGDZVP basis is NOT the member of Dunning's cc-type basis set
family. Rather, it's the basis set used by DGauss and is designed for
DFT calculations. It is availabe at:

http://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtml

but unfortunately in DGauss format. This format is documented
inside the documents containing basis sets themselves. Note you
do not need to worry on auxiliary (fitting) basis sets at all.

You can write some script to format it as PC GAMESS/Firefly
external basis set file. See the cc-type basis sets available in the
"Downloads" section of our web server for samples of the external
basis set files. Please share it with us if you really create GAMESS
format external basis set library for DGDZVP.

2. The PC GAMESS does not store Dunning's cc-type basis
sets internally, however, they can be used using

 $basis extfil=.t. gbasis=name $end

directive. You may need to use -b command line option to set the
path to and the name of the basis set library file. There are
multiple files containng definitions of popular external basis
sets available in the "Downloads" section.

3. The GAMESS' ispher option is not supported, however,
the PC GAMESS-specific d5 option and related $d5 group
can be used instead. Together, they allow finer control on removal of
spherical contaminants. This option is documented in the "Manuals"
section:


http://classic.chem.msu.su/gran/gamess/spherical.html

4. Finally, if you use external basis sets, any of NDFUNC, etc... are ignored.

Regards,
Alex



[ This message was edited on Tue Nov 18 '08 at 4:05pm by the author ]


Message 4df668afbHW-6897-995+00.htm, number 639, was posted on Tue Nov 18 '08 at 16:34:56
in reply to dcd10513NCe-6896-1021+00.htm

Re^3: Geometry optimization of S1 state by MCSCF

Alex Granovsky
gran@classic.chem.msu.su


Dear Prof. Mohri,

>>! Need extra states in CI as the second one is triplet
>> $DET NCORE=7 NACT=2 NELS=2 NSTATE=4 GROUP=C1 ITERMX=200
>>PURES=.T. $END
> I understand that PURES=.T. means that only singlet states are included in the
>state-averaging. Is this understanding correct?

Indeed, PURES=.T. means that only the states with spin
equal to that specified in input file will be included in
state-averaging procedure. Note that pures=.t. is the default
and strictly speaking it was not necessary to use this option.
Another point is that the default value of SZ is set according
to the MULT value in $contrl. However, this is the aldet CI code,
and thus it works with determinants rather than CSF.

This means it solves for NSTATE low-liyng roots regardless
of their multiplicity. The roots will be renumbered at the
latter states for the purpose of state-averaging. Thus, the
second CI root is triplet, however, the second root in
averaging is indeed singlet (S1 state). This S1 state is
the third root of determinant CI.

> The sentence of 'the second one is
> triplet' makes me confused.
> Why the triplet state appears in this case?

Because this is determinant CI code. You can use GUGA
code, which works with CFS and will result in states
of only selected multiplicity. However, for large
active spaces ALDET is usually faster (well, the sample
itself is definitely not the large problem).

> Moreover,
> how did you determine the number of states of four, that is,
> how did you determine
> NSTATE=4 before doing this run?
The answer is that for ALDET one usually needs some extra
roots to capture S1 (or SN). In this particular case of CAS (2,2)
it is evident that there exist only four states at all.


>! State-averaged MCSCF
> $DET WSTATE(1)=1,1 $END
>! istate=2 means we need properties and gradiends of second
> state (S1)
> $mcscf CISTEP=aldet maxit=200 ISTATE=2 $end
> The expression of 'second state (S1)' also makes me confused,
> because this seems to be inconsistent with the above sentence
> 'the second one is triplet'. I imagine that the triplet state is
> removed and thus the S1 state is 2nd lower state among all singlet
> states. Is this correct?

Exactly. With default pures settings, the CI roots are
renumbered so that the states of wrong spin are not counted.


> In opt.s1.sa.out
> STATE 2 ENERGY= -113.6672881517 S= 1.00 SZ= 0.00 SPACE SYM=A

>ALPH|BETA| COEFFICIENT
>----|----|------------
> 01 | 10 | -0.7071068
> 10 | 01 | 0.7071068
> 10 | 10 | 0.0000000
>
> I think that the above combinations of ALPHA and BETA
> should be of singlet states. But S= 1.00 is output. Why?
> (I think that this means 2S+1=3)

This is the triplet state (exactly the same I meant while
mention that the second state is triplet). More precisely,
it's the triplet state having SZ=0.

Regards,
Alex Granovsky


Message 4df668afbHW-6897-1019+00.htm, number 639, was edited on Tue Nov 18 '08 at 16:59:06
and replaces message 4df668afbHW-6897-995+00.htm

Re^3: Geometry optimization of S1 state by MCSCF

Alex Granovsky
gran@classic.chem.msu.su


Dear Prof. Mohri,

>>! Need extra states in CI as the second one is triplet
>> $DET NCORE=7 NACT=2 NELS=2 NSTATE=4 GROUP=C1 ITERMX=200
>>PURES=.T. $END
> I understand that PURES=.T. means that only singlet states are included in the
>state-averaging. Is this understanding correct?

Indeed, PURES=.T. means that only the states with spin
equal to that specified in input file will be included in
state-averaging procedure. Note that pures=.t. is the default
and strictly speaking it was not necessary to use this option.
Another point is that the default value of SZ is set according
to the MULT value in $contrl. However, this is the aldet CI code,
and thus it works with determinants rather than CSF.

This means it solves for NSTATE low-liyng roots regardless
of their multiplicity. The roots will be renumbered at the
subsequent stages for the purpose of state-averaging. Thus,
the second CI root is triplet, however, the second root in
averaging is indeed singlet (S1 state). This S1 state is
the third root of determinant CI.

> The sentence of 'the second one is
> triplet' makes me confused.
> Why the triplet state appears in this case?

Because this is determinant CI code. You can use GUGA
code, which works with CFS and will result in states
of only selected multiplicity. However, for large
active spaces ALDET is usually faster (well, the sample
itself is definitely not the large problem).

> Moreover,
> how did you determine the number of states of four, that is,
> how did you determine
> NSTATE=4 before doing this run?
The answer is that for ALDET one usually needs some extra
roots to capture S1 (or SN). In this particular case of CAS (2,2)
it is evident that there exist only four states at all.


>! State-averaged MCSCF
> $DET WSTATE(1)=1,1 $END
>! istate=2 means we need properties and gradiends of second
> state (S1)
> $mcscf CISTEP=aldet maxit=200 ISTATE=2 $end
> The expression of 'second state (S1)' also makes me confused,
> because this seems to be inconsistent with the above sentence
> 'the second one is triplet'. I imagine that the triplet state is
> removed and thus the S1 state is 2nd lower state among all singlet
> states. Is this correct?

Exactly. With default pures settings, the CI roots are
renumbered so that the states of wrong spin are not counted.


> In opt.s1.sa.out
> STATE 2 ENERGY= -113.6672881517 S= 1.00 SZ= 0.00 SPACE SYM=A

>ALPH|BETA| COEFFICIENT
>----|----|------------
> 01 | 10 | -0.7071068
> 10 | 01 | 0.7071068
> 10 | 10 | 0.0000000
>
> I think that the above combinations of ALPHA and BETA
> should be of singlet states. But S= 1.00 is output. Why?
> (I think that this means 2S+1=3)

This is the triplet state (exactly the same I meant while
mention that the second state is triplet). More precisely,
it's the triplet state having SZ=0.

Regards,
Alex Granovsky

[ This message was edited on Tue Nov 18 '08 at 4:59pm by the author ]


Message 4df668afbHW-6897-1020+00.htm, number 639, was edited on Tue Nov 18 '08 at 17:00:28
and replaces message 4df668afbHW-6897-1019+00.htm

Re^3: Geometry optimization of S1 state by MCSCF

Alex Granovsky
gran@classic.chem.msu.su


Dear Prof. Mohri,

>>! Need extra states in CI as the second one is triplet
>> $DET NCORE=7 NACT=2 NELS=2 NSTATE=4 GROUP=C1 ITERMX=200
>>PURES=.T. $END
> I understand that PURES=.T. means that only singlet states are included in the
>state-averaging. Is this understanding correct?

Indeed, PURES=.T. means that only the states with spin
equal to that specified in input file will be included in
state-averaging procedure. Note that pures=.t. is the default
and strictly speaking it was not necessary to use this option.
Another point is that the default value of SZ is set according
to the MULT value in $contrl. However, this is the aldet CI code,
and thus it works with determinants rather than CSF.

This means it solves for NSTATE low-liyng roots regardless
of their multiplicity. The roots will be renumbered at the
subsequent stages for the purpose of state-averaging. Thus,
the second CI root is triplet, however, the second root in
averaging is indeed singlet (S1 state). This S1 state is
the third root of determinant CI.

> The sentence of 'the second one is
> triplet' makes me confused.
> Why the triplet state appears in this case?

Because this is determinant CI code. You can use GUGA
code, which works with CSF and will result in states
of only selected multiplicity. However, for large
active spaces ALDET is usually faster (well, the sample
itself is definitely not the large problem).

> Moreover,
> how did you determine the number of states of four, that is,
> how did you determine
> NSTATE=4 before doing this run?
The answer is that for ALDET one usually needs some extra
roots to capture S1 (or SN). In this particular case of CAS (2,2)
it is evident that there exist only four states at all.


>! State-averaged MCSCF
> $DET WSTATE(1)=1,1 $END
>! istate=2 means we need properties and gradiends of second
> state (S1)
> $mcscf CISTEP=aldet maxit=200 ISTATE=2 $end
> The expression of 'second state (S1)' also makes me confused,
> because this seems to be inconsistent with the above sentence
> 'the second one is triplet'. I imagine that the triplet state is
> removed and thus the S1 state is 2nd lower state among all singlet
> states. Is this correct?

Exactly. With default pures settings, the CI roots are
renumbered so that the states of wrong spin are not counted.


> In opt.s1.sa.out
> STATE 2 ENERGY= -113.6672881517 S= 1.00 SZ= 0.00 SPACE SYM=A

>ALPH|BETA| COEFFICIENT
>----|----|------------
> 01 | 10 | -0.7071068
> 10 | 01 | 0.7071068
> 10 | 10 | 0.0000000
>
> I think that the above combinations of ALPHA and BETA
> should be of singlet states. But S= 1.00 is output. Why?
> (I think that this means 2S+1=3)

This is the triplet state (exactly the same I meant while
mention that the second state is triplet). More precisely,
it's the triplet state having SZ=0.

Regards,
Alex Granovsky

[ This message was edited on Tue Nov 18 '08 at 5:00pm by the author ]


Message d8ed14aa6g1-6897-1181+00.htm, number 640, was posted on Tue Nov 18 '08 at 19:41:35
in reply to 4df668afbHW-6897-965+00.htm

Re^4: B3YLP5 Job with Dunning Basis Set

Thomas Patko
tpatko@gmail.com


On Tue Nov 18 '08 4:05pm, Alex Granovsky wrote
----------------------------------------------
>Hello Thomas,

>I hope the following comments would be helpful:

>1. The DGDZVP basis is NOT the member of Dunning's cc-type basis set
>family. Rather, it's the basis set used by DGauss and is designed for
>DFT calculations. It is availabe at:

>

http://www.ccl.net/cca/data/basis-sets/DGauss/dgauss-basis.shtml

>but unfortunately in DGauss format. This format is documented
>inside the documents containing basis sets themselves. Note you
>do not need to worry on auxiliary (fitting) basis sets at all.

>You can write some script to format it as PC GAMESS/Firefly
>external basis set file. See the cc-type basis sets available in the
>"Downloads" section of our web server for samples of the external
>basis set files. Please share it with us if you really create GAMESS
>format external basis set library for DGDZVP.

OK. I will give it a try and notify you if I get anything working for PC-GAMESS.

>2. The PC GAMESS does not store Dunning's cc-type basis
>sets internally, however, they can be used using

>

 $basis extfil=.t. gbasis=name $end

>directive. You may need to use -b command line option to set the
>path to and the name of the basis set library file. There are
>multiple files containng definitions of popular external basis
>sets available in the "Downloads" section.

OK. Is there any plans to have the Dunning CC and aug-cc basis sets stored internally in the next release of PCG?

>3. The GAMESS' ispher option is not supported, however,
>the PC GAMESS-specific d5 option and related $d5 group
>can be used instead. Together, they allow finer control on removal of
>spherical contaminants. This option is documented in the "Manuals"
>section:
>
>
>

http://classic.chem.msu.su/gran/gamess/spherical.html

>4. Finally, if you use external basis sets, any of NDFUNC, etc... are ignored.

>Regards,
>Alex


Message 829e844cbK2-6898-746-00.htm, number 641, was posted on Wed Nov 19 '08 at 12:26:27
calculation failed for rdmini

Mahesh Bhatt
mbhatt54@yahoo.com


Dear Sir,
as you suggested in attached input file for rdmini for bcp in last October, I have tried to make calculation. But it terminates abnormally. My input is as follows:
!*** PC GAMESS INPUT generated by Facio 11.4.1 ***
$CONTRL ICHARG=0 MULT=1 SCFTYP=RHF RUNTYP=ENERGY LOCAL=NONE
COORD=UNIQUE NZVAR=0 MAXIT=300 NPRINT=7 NOSYM=0
PLTORB=.TRUE. EXETYP=RUN IREST=0 DFTTYP=B3LYP5
$END
$SYSTEM TIMLIM=600 MEMORY=90000000 $END
$SCF dirscf = 1 soscf = 1 $END
$BASIS GBASIS=N31 NGAUSS=6
NDFUNC=1 NFFUNC=0 NPFUNC=1 DIFFSP=.FALSE. DIFFS=.FALSE.
$END
$GUESS GUESS=rdmini $END
$ELDENS IEDEN=1 ICORBS(1)=225, 226 $END
$ELPOT IEPOT=0 $END
$CUBE CUBE=.TRUE. MESH=MEDIUM $END
$soscf crit = 1000000 $end
$DATA
*** \GAMESS\pcgamess_windows_p5\Facio\PDB1\Ca13(111) surface.pdb ***
C1 0
Ca 20.0 0.000 0.000 0.000
Ca 20.0 3.950 0.000 0.000
Ca 20.0 -3.950 0.000 0.000
Ca 20.0 7.900 0.000 0.000
Ca 20.0 -7.900 0.000 0.000
Ca 20.0 1.980 3.940 0.000
Ca 20.0 -1.970 3.950 0.000
Ca 20.0 -5.920 3.950 0.000
Ca 20.0 5.930 3.950 0.000
CA 20.0 1.980 -3.950 0.000
CA 20.0 5.930 -3.950 0.000
CA 20.0 -1.970 -3.950 0.000
Ca 20.0 -5.920 -3.950 0.000
C 6.0 -2.828 3.588 4.192
C 6.0 -1.741 2.734 4.071
C 6.0 -1.847 1.383 4.430
C 6.0 -3.072 0.917 4.916
C 6.0 -4.162 1.773 5.031
C 6.0 -4.045 3.109 4.665
H 1.0 -2.727 4.645 3.925
H 1.0 -0.771 3.116 3.720
H 1.0 -5.111 1.394 5.420
H 1.0 -4.907 3.780 4.752
C 6.0 0.579 1.151 4.566
C 6.0 1.770 0.397 4.618
H 1.0 0.621 2.246 4.676
C 6.0 3.082 1.076 4.816
H 1.0 3.820 0.435 5.317
H 1.0 2.965 1.985 5.422
H 1.0 3.510 1.375 3.848
N 7.0 1.746 -0.940 4.513
C 6.0 -0.646 0.531 4.372
C 6.0 1.733 -5.051 4.621
C 6.0 0.532 -5.789 4.570
C 6.0 -0.685 -5.155 4.376
C 6.0 3.036 -5.748 4.822
H 1.0 0.561 -6.885 4.680
H 1.0 2.905 -6.655 5.427
H 1.0 3.781 -5.116 5.323
H 1.0 3.460 -6.053 3.854
N 7.0 1.727 -3.715 4.515
C 6.0 -3.118 -5.507 4.906
C 6.0 -1.896 -5.991 4.432
C 6.0 -1.806 -7.346 4.082
C 6.0 -2.904 -8.186 4.202
C 6.0 -4.117 -7.687 4.667
C 6.0 -4.218 -6.348 5.022
H 1.0 -0.837 -7.742 3.741
H 1.0 -2.815 -9.246 3.944
H 1.0 -4.988 -8.346 4.756
H 1.0 -5.162 -5.951 5.409
H 1.0 -3.177 -0.140 5.195
H 1.0 -3.212 -4.447 5.181
C 6.0 -0.661 -0.880 4.184
C 6.0 0.553 -1.591 4.319
C 6.0 0.543 -3.049 4.321
C 6.0 -0.680 -3.744 4.187
C 6.0 -1.851 -2.982 3.866
C 6.0 -1.842 -1.627 3.865
H 1.0 -2.776 -3.530 3.637
H 1.0 -2.760 -1.070 3.632
$END
When I applied this input and tried to make calculation, there was a message in Facio window as:

GAMESS Energy Calculation has started.
*** Please wait patiently. ***
Calculation has finished.
Execution of GAMESS terminated abnormally
*** PC GAMESS Output***

I would like to request you that previous input options for scf convergence was OK. Now Can I use this convergence input option to calculate rdmini?
i am waiting for your reply.
Sincerely
Mahesh Bhatt
University of Tsukuba
Japan


Message 4df668afbHW-6898-925+00.htm, number 642, was posted on Wed Nov 19 '08 at 15:25:15
in reply to d8ed14aa6g1-6897-1181+00.htm

Re^5: B3YLP5 Job with Dunning Basis Set

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>OK. Is there any plans to have the Dunning CC and aug-cc basis sets stored internally in the next release of PCG?

Frankly, it is unlikely. The reason is that
the externally stored basis sets allow much
more flexibility and can be easily corrected
in the case of any errors/misprints, etc...
More likely, we will change the way how basis
sets are specified in the input file to allow
simpler, more convenient and more user-friendly
specification.

Regards,
Alex


Message 4df668afbHW-6898-926+00.htm, number 624, was edited on Wed Nov 19 '08 at 15:26:21
and replaces message 4df668afbHW-6895-1195+00.htm

Re^3: IRC computations with Firefly

Alex Granovsky
gran@classic.chem.msu.su


Hi,

>That is exactly the point which I would like to address. Sometimes one faces flat (i.e. small curvature) PES and therefore one needs to choose *larger* STRIDE values. At the same time, if one wants to avoid bad stops due to the flatness of the IRC, one should set the OPTTOL threshold to *smaller* values. Therefore IMHO the OPTTOL value should also be tunable by the Firefly users.

The opttol variable is tunable. The point is that
(as I already explained) it usually has very little effect on IRC calculations.

On a separat note, are you sure you are really interested
in changing the opttol value? rather than just the values of
rcut/gcut parameters?

Regards,
Alex

[ This message was edited on Wed Nov 19 '08 at 3:26pm by the author ]


Message d4c0ee0abHW-6898-934+00.htm, number 616, was edited on Wed Nov 19 '08 at 15:34:12
and replaces message d4c0ee0abHW-6892-1371+00.htm

Re^7: PC GAMESS under Linux Ubuntu 64bits: dynamically linked version works normally

Alex Granovsky
gran@classic.chem.msu.su


Hi,

I checked the output of strace you sent to me and it seems that the problem is with the


sendmsg()

and

recvmsg()

calls - they seem to be improperly handled by the 32-bit subsystem
of your 64-bit Linux version.
You may file bug report to developers if you like :-)

Regards,
Alex Granovsky

[ This message was edited on Wed Nov 19 '08 at 3:34pm by the author ]


Message 4405ef556g1-6898-1075+00.htm, number 643, was posted on Wed Nov 19 '08 at 17:55:29
in reply to 4df668afbHW-6898-925+00.htm

Re^6: B3YLP5 Job with Dunning Basis Set

Thomas Patko
tpatko@gmail.com


Does anyone have experience running a B3YLP DFT jobs with a basis set that support iodine on PC-GAMESS? Essentially I will need a basis set that supports the main group elements through Xenon rather than just Krypton. I would need to be a double zeta with polarization (essentially akin to 6-31G* but with support for the next row of main group elements). If only triple zeta basis sets offer this support I could do that as well, it just might slow down my computations and probably not improve my results significantly (if at all). If such a basis set if correlation consistent then all the better.

Any experience or thoughts are appreciated. If such a basis i set is already available from the current PCG downloads section (rather than building it from the Basis Set Exchange website (https://bse.pnl.gov/bse/portal) that would be good as well.

Cheers,

Thomas


Message 4405ef556g1-6898-1089+00.htm, number 644, was posted on Wed Nov 19 '08 at 18:09:17
in reply to 4df668afbHW-6898-925+00.htm

Re^6: B3YLP5 Job with Dunning Basis Set

Thomas Patko
tpatko@gmail.com


Hello Alex:

I can see the advantages of external basis sets. Do you envision something like a system were PCG checks if the basis set called is available internally, and if not available internally then checks a default subdirectory path (maybe ./basis/) where well tested external basis sets can be provided in the default PCG distribution? A third tier of basis sets could be those that are not stored internally and not distribution with the main PCG package (not well tested enough, not yet available, novel , experimental, ...).

Cheers,

TADP

On Wed Nov 19 '08 3:25pm, Alex Granovsky wrote
----------------------------------------------
>Hi,

>>OK. Is there any plans to have the Dunning CC and aug-cc basis sets stored internally in the next release of PCG?

>Frankly, it is unlikely. The reason is that
>the externally stored basis sets allow much
>more flexibility and can be easily corrected
>in the case of any errors/misprints, etc...
>More likely, we will change the way how basis
>sets are specified in the input file to allow
>simpler, more convenient and more user-friendly
>specification.

>Regards,
>Alex
>


Message 51acec17RPr-6899-10-00.htm, number 645, was posted on Thu Nov 20 '08 at 00:12:20
Calculating the Hessian

amir barazande
barazandeamir@gmail.com


Hi there!
I'm a newbe in the field, and i want to do a Hessian on Benzene.
But i get some kind of warning message from GAMESS in the out-file as below;
*******************************************************
* THIS IS NOT A STATIONARY POINT ON THE MOLECULAR PES *
* THE VIBRATIONAL ANALYSIS IS NOT VALID !!! *
*******************************************************
Could anybody give me a hint?
What do i do wrong(?)
How can i modify my input file in order to get rid of this message?

Here is my inp-file content:
-----------------------------------------------------------

$CONTRL SCFTYP=RHF RUNTYP=HESSIAN $END
$CONTRL EXETYP = RUN MOLPLT=.TRUE. $END
$BASIS GBASIS=STO NGAUSS=3 $END
$DATA
BenxeneHessian
C1
C 6.0 -0.50394 0.12775 -0.07763
C 6.0 0.70416 -0.56975 -0.07763
C 6.0 0.70416 -1.96475 -0.07763
C 6.0 -0.50394 -2.66225 -0.07763
C 6.0 -1.71204 -1.96475 -0.07763
C 6.0 -1.71204 -0.56975 -0.07763
H 1.0 -0.50394 1.22745 -0.07763
H 1.0 1.65656 -0.01985 -0.07763
H 1.0 1.65656 -2.51465 -0.07763
H 1.0 -0.50394 -3.76195 -0.07763
H 1.0 -2.66444 -2.51465 -0.07763
H 1.0 -2.66444 -0.01985 -0.07763
$END
$SCF DIRSCF=.TRUE. $END
$STATPT NSTEP=200 $END
$FORCE METHOD=ANALYTIC $END
-----------------------------------------------
Thanks in advance
/Amir

This message contains the 103 kb attachment
[ hessianBenz.out ] Here is the output-file


Message 558d7095Uee-6899-259+00.htm, number 646, was posted on Thu Nov 20 '08 at 04:19:21
in reply to 51acec17RPr-6899-10-00.htm

Re: Calculating the Hessian

sanya
sanya@photonics.ru


> *******************************************************
> * THIS IS NOT A STATIONARY POINT ON THE MOLECULAR PES *
> * THE VIBRATIONAL ANALYSIS IS NOT VALID !!! *
> *******************************************************

It's easy. You first have to do geometry optimization ($CONTRL runtyp=optimize $end). Otherwise, you'll get meaningless vibrational frequencies. If you've got this message after geometry optimization, try to set lower geometry optimization threshold ($statpt opttol=1e-5 $end), re-optimize the geometry and try hessian again.


Message 558d7095Uee-6899-272+00.htm, number 647, was posted on Thu Nov 20 '08 at 04:32:54
in reply to 4405ef556g1-6898-1075+00.htm

Re^7: B3YLP5 Job with Dunning Basis Set

sanya
sanya@photonics.ru


Why not using https://bse.pnl.gov/bse/portal? It's not so inconvenient :) But be careful about what is discussed here http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C35b4c45d5Uee-6885-1414+00.htm.

Here is the list of basis sets supported for iodine (from http://www.emsl.pnl.gov/forms/basisform-orig.html:

STO-3G
3-21G
6-311G
6-311G*
6-311G**
MIDI!
cc-pVDZ-PP
cc-pVTZ-PP
cc-pVQZ-PP
cc-pV5Z-PP
aug-cc-pVDZ-PP
aug-cc-pVTZ-PP
aug-cc-pVQZ-PP
aug-cc-pV5Z-PP
WTBS
Sadlej pVTZ
LANL2DZ ECP
LANL2DZdp ECP
SBKJC VDZ ECP
CRENBL ECP
Stuttgart RLC ECP
SDB-cc-pVTZ
SDB-cc-pVQZ
SDB-aug-cc-pVTZ
SDB-aug-cc-pVQZ
DZVP (DFT Orbital)
DeMon Coulomb Fitting
DGauss A1 DFT Coulomb Fitting
DGauss A1 DFT Exchange Fitting
Ahlrichs Coulomb Fitting
6-311G* Polarization
6-311G** Polarization
Huzinaga Polarization
SDB-aug-cc-pVTZ Diffuse
SDB-aug-