Firefly and PC GAMESS-related discussion club

 
Learn how to ask questions correctly  
 
 

 

Alex Granovsky
gran@classic.chem.msu.su

>Thank you for a detailed information. I have a doubt regarding the $VIB group...I checked the o/p file, PUNCH file and the IRCDATA file..All of them contain $VIB.. Whats the difference between all of them?

Normally, you need a $VIB group from the IRCDATA file. Other $VIB
groups are printed for mainly for backup purposes and have their
data scattered across entire output and punch files.

>All I have is the NSERCH value..Say NSERCH in the o/p file was "0", corresponding to the lowest energy.. So now in the PUNCH file, is it fine if I search for NSERCH=0? and the coordinates following it?

Yes, it is fine to search punch file for NSERCH value.

> ALSO, what is the difference between "FINAL ENERGY" and the ENERGY I see alongwith NSERCH=NNN ENERGY=EEE string?

The FINAL ENERGY data are printed by SCF routines. For plain SCF,
these values are the same, as the energy you see alongwith NSERCH=NNN
ENERGY=EEE string. For SA-MCSCF, any CI, TDHF/TDDFT, and any MPn
(MP2, MP3, MP4,etc...), they are not the same.

A couple of additional notes:

If you need to restart Saddle point location run,
you should restart it from the search point having lowest RMS
or MAXIMUM gradient values rather than lowest energy value.
For this, you can search output for "RMS GRADIENT" string.

If you want to be able to restart Hessian jobs, you should always
use COORD=UNIQUE, throughout initial and all subsequent jobs.
The reason is that this type of coordinates does not reorient
input geometry leaving it intact. It is good idea to always
use COORD=UNIQUE.

Kind regards,
Alex Granovsky

On Tue May 21 '13 9:17am, Siddheshwar Chopra wrote
--------------------------------------------------
>Dear Prof. Alex,
>Thank you for a detailed information. I have a doubt regarding the $VIB group...I checked the o/p file, PUNCH file and the IRCDATA file..All of them contain $VIB.. Whats the difference between all of them? Also when you told about searching for the string "NSERCH".. I get the lowest energy..But when I go to the PUNCH file, I don't see this lowest energy value there..Alll I have is the NSERCH value..Say NSERCH in the o/p file was "0", corresponding to the lowest energy.. So now in the PUNCH file, is it fine if I search for NSERCH=0? and the coordinates following it? ALSO, what is the difference between "FINAL ENERGY" and the ENERGY I see alongwith NSERCH=NNN ENERGY=EEE string?

>Regards,

>On Mon May 20 '13 6:32pm, Alex Granovsky wrote
>----------------------------------------------
>>Hello,

>>In addition, the same way can be used to restart Raman activity
>>calculations. More precisely, one need to copy a first $vec
>>group from the punch file and a $vib group from ircdata to the
>>input file (do not forget to properly terminate $vib group by
>>$end directive); and add $guess guess=moread $end to the input.

>>Kind regards,
>>Alex Granovsky
>>
>>
>>On Mon May 20 '13 6:25pm, Alex Granovsky wrote
>>----------------------------------------------
>>>Dear Siddheshwar,

>>>You can find information on how to restart hessian runs by exploring this thread:

>>>http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C3d8ed14aa6g1-7050-1262+00.htm

>>>Hope this helps.

>>>Kind regards,
>>>Alex Granovsky
>>>
>>>
>>>On Mon May 20 '13 2:20am, Alex Granovsky wrote
>>>----------------------------------------------
>>>>Dear Siddheshwar,

>>>>One of the ways to restart geometry optimization run is to
>>>>search output file for the geometry having the lowest energy,
>>>>looking for strings like:

>>>>          NSERCH=  NNN     ENERGY=    EEE

>>>>Here NNN is a step number and EEE is an energy. The simplest way
>>>>to do this is to use "grep" (under Linux) or "find" (under Windows)
>>>>command line tools Note, grep is very useful and much more
>>>>powerful tool. There are many ports of grep for Windows, so you
>>>>can download and try one of them. There also exists a SUA
>>>>(Subsystem for Unix Applications) subsystems for Windows which
>>>>is available from Microsoft.

>>>>Upon locating the geometry having the lowest energy,
>>>>create a copy of your input file and replace initial
>>>>coordinates by the coordinates that correspond to the
>>>>lowest energy (these coordinates can be found in the output
>>>>a bit above the string with the lowest energy data).

>>>>Next, look at the punch file. It duplicates the most important
>>>>parts of output and contains some additional information.
>>>>For instance, it has a set formatted $vec groups that
>>>>contains MOs at all geometry optimization points. Copy
>>>>the $vec group corresponding to NSERCH with minimum energy
>>>>to the new input file and use

>>>>

 $guess guess=moread norb=nnnn $end

>>>>You should replace nnnn by the actual number of MOs
>>>>(you should provide at least all populated orbitals).

>>>>That's basically all. I'll describe other ways of how to
>>>>restart optimization jobs in my next posts and also explain
>>>>how to restart hessian calculations.

>>>>Hope this helps.

>>>>Kind regards,
>>>>Alex Granovsky
>>>>
>>>>
>>>>On Thu May 16 '13 11:53am, Siddheshwar Chopra wrote
>>>>---------------------------------------------------
>>>>>Dear Alex,
>>>>>My jobs particularly involve geometry optimization, numerical hessian jobs, finding raman, using DFT, TDFT , mostly. Sir, I need to know the code which I have to add to my Restart i/p file, WHICH WOULD START THE CALCULATION FROM WHERE IT WAS CRASHED (OR STOPPED).. Please make me understand this well. I would be grateful if you could send me an i/p file or the commands..

>>>>>Regards,
>>>>>
>>>>>
>>>>>On Wed May 15 '13 9:55pm, Alex Granovsky wrote
>>>>>----------------------------------------------
>>>>>>Hello,
>>>>>>
>>>>>>
>>>>>>It is rather straightforward to restart geometry optimization
>>>>>>and numerical hessian jobs. Are these the types of jobs you
>>>>>>are interested in? Please let us know.

>>>>>>Several other types of jobs can be restarted as well but it is more tricky.

>>>>>>A good solution would be to buy an UPS. It would also help protecting
>>>>>>your hardware from damage due to power outages, electrical surges,  noise etc...

>>>>>>Kind regards,
>>>>>>Alex Granovsky
>>>>>>
>>>>>>
>>>>>>On Wed May 15 '13 12:04pm, Siddheshwar Chopra wrote
>>>>>>---------------------------------------------------
>>>>>>>Dear Sir,
>>>>>>>Due to erratic power cuts or other reasons, it is becoming very difficult to restart any calculations everytime.. PLEASE TELL ME is there any way I can RESTART from the same, once the FIREFLY crashes due to such power issue?? It would be a great help appreciated in advance.

>>>>>>>Regards,