Page 1 of 1

how to skip SCF for response calculation?

Posted: 03 Jun 2014, 09:49
by bogdan
hi,

is it possible to skip completely the SCF part for a response calculation? i'm looking for a keyword similar to the .RESTART one in the *DENSOPT section,
in the sense that it reads the available dens.restart file, but from that moment on, the program should assume that the SCF is already converged.
in gamess one can achieve this by specifying scftyp=none and then read a $VEC file with guess=moread option.

thanx,
bogdan

Re: how to skip SCF for response calculation?

Posted: 03 Jun 2014, 10:34
by tkjaer
I am confused - why do you not just use

*DENSOPT
.RESTART

it will restart SCF - do a single iteration to verify that it is converged and start the response.

TK

Re: how to skip SCF for response calculation?

Posted: 03 Jun 2014, 10:34
by tkjaer
it does require a dens.restart file in the directory where LSDALTON.INP and MOLECULE.INP is placed.

Re: how to skip SCF for response calculation?

Posted: 03 Jun 2014, 11:42
by bogdan
tkjaer wrote:I am confused - why do you not just use

*DENSOPT
.RESTART

it will restart SCF - do a single iteration to verify that it is converged and start the response.

TK
Thanx Thomas,

at the moment i'm using the .RESTART feature but the problem is that one SCF iteration takes around 2-3 hours for the calculations i'm currently running. and i don't want to waste this time since the maximum allowed job time can't exceed 24 hours.

/bogdan

Re: how to skip SCF for response calculation?

Posted: 03 Jun 2014, 11:46
by tkjaer
There is no such option. I could add it for the next release but currently not such option exist.

TK

Re: how to skip SCF for response calculation?

Posted: 03 Jun 2014, 11:52
by bogdan
Thanx alot,

if that doesn't require a considerable programming effort i think implementing this would be quite nice.

/bogdan

Re: how to skip SCF for response calculation?

Posted: 04 Jun 2014, 11:48
by tkjaer
the keyword

.SKIPSCFLOOP

will now be included in the Dalton2014 release