Firefly and PC GAMESS-related discussion club


 
Learn how to ask questions correctly  
 
 
We are NATO-free zone
 



Re: Crash: ATTEMPT TO READ A DAF RECORD THAT WAS NEVER WRITTEN

Alex Granovsky
gran@classic.chem.msu.su


Dear Bernhard,

your input file is incorrect and this causes error during execution.

Your input file contains multiple $vec groups and multiple $grad
groups. I suppose this is a copy-paste error.

Valid input file should only contain a single $vec group if any.
In addition, input file for conical intersection location runs
(i.e. those with method=conic in $statpt group) should not contain
any $grad groups. Firefly runs just fine when these input errors
are fixed. I'll add an explicit check for conic runs with $grad
group specified in the input file to trap these runs at the earlier
stages.

On a separate note, if you replace the following line:

 $TRANS CUTTRF=1D-12 DIRTRF=.T. $END

in your input file by this one:

 $TRANS CUTTRF=1D-12 DIRTRF=.T. MPTRAN=2 MODE=12 altpar=1 $END

Firefly should work several times faster.
Note, you cannot use MPTRAN=2 code with fullnr converger,
it is designed specifically for soscf and focas convergers.

Kind regards,
Alex Granovsky








On Fri Mar 31 '17 3:52pm, Bernhard Dick wrote
---------------------------------------------
>I encounter a stop of firefly with the error message:

> *** ERROR *** ATTEMPT TO READ A DAF RECORD THAT WAS NEVER WRITTEN.
> RECORD NUMBER  157 OF LENGTH       132 DOES NOT EXIST.

> ADDRESS 0x005AFBBB HAS INITIATED PROGRAM ABORT BECAUSE OF FATAL ERROR(S)

>The error occurs near the calculation of state specific gradients from a state-averaged CAS calculation, in the context of finding a conical intersection. I attach the input and output files. The error is reproducible and occurs under Windows and also on a Linux cluster. However, it does not occur with exactly the same type of calculation for a somewhat different molecule of similar size.

>with best regards,
>Bernhard


[ Previous ] [ Next ] [ Index ]           Thu Apr 6 '17 9:29pm
[ Reply ] [ Edit ] [ Delete ]           This message read 239 times