Help with large CCSD(T) Calculation


Click here for full thread
Gets Around
Quote:Edoapra Aug 23rd 9:22 pm
Keithusin
The input you are using can be run with the TCE module if you increase the number of processors.
If you want to stick to 16 processors, you might want to switch to the older "CCSD" module that
has small memory requirements.
I have been trying to reproduce the behavior of the input you are using and
I have come up with the two input files below.
Please keep in mind that because of the different memory requirements
for CCSD and the (T) part, you will have to use two different input files
 
start ccsd
title "CCSD input"
memory stack 800 mb heap 100 mb global 750 mb
geometry
        C   2.12544   0.00000   0.00000
        O   1.82852  -0.93172  -0.62769
        O   2.42235   0.93172   0.62769
        C  -2.12544   0.00000   0.00000
        O  -1.20623  -0.32695  -0.63119
        O  -3.04465   0.32695   0.63119
end
basis
* library aug-cc-pvqz
end

scf
 direct
 thresh 1d-8
end

ccsd
 diisbas 2
 freeze atomic
 nodisk
 tol2e 1d-14
end

task ccsd


restart ccsd

title "CCSD(T) input"

memory stack 400 mb heap 100 mb global 950 mb

task ccsd(t)


Hello Edoapra,
I just need to understand something..
From this code of yours how can you be able to get the interaction energy term between CO2--CO2.
I mean which portion of this code is defining that..??