Firefly and PC GAMESS-related discussion club



Learn how to ask questions correctly


Re^3: Speeding-up optimization and vibrational analysis

sanya
sanya@photonics.ru


>As you will see in the attachment there is a message concerning linear dependence.

Remove all the diffuse basis functions. You don't need them for frequency calculation. Alternatively, you may try to remove them step-by-step as stated in the message. In this case, you'll have to give basis set individually for each atom (see forum below).

>As for a direct calculation, I think that the calculation time will increase dramatically.

That's not true. Just try and see that it will speed up your calculation. All tasks with more than 50 basis functions should be run in DIRSCF mode.

>I shall try another process without HESS=CALC. It will take, probably, several days on my computer (2.4 GHz Intel dual processor T8300. The firefly implementation on a linux cluster is still pending)
>Please have a look at the log file and give me a hint.

This task can be easily done on a single node on two cores (you run it on one core, so check your command line).

Add the following keywords to your $CONTRL group

 fstint=.t. gencon=.t. d5=.t.

and the following groups

$smp load=0 call64=.t. $end
$p2p p2p=.t. dlb=.t. $end
$trans mptran=2 dirtrf=.t. aoints=dist altpar=.t. mode=112 $end

Anyway, if they are not applicable to your hardware/software, theyl'll just be ignored.


[ Previous ] [ Next ] [ Index ]           Fri Dec 17 '10 2:18am
[ Reply ] [ Edit ] [ Delete ]           This message read 876 times