cphf solve3 issues

sophia sagala

unread,

Feb 4, 2021, 3:02:47 AM2/4/21

to nwchem...@googlegroups.com

Hello

I am currently trying to perform a vibrational calculation on a molecule to obtain Raman active wavenumber and intensities. I have been receiving the error of,

cphf_solve3:SCF residual greater than 1d-2

shortly after the CPHF module is called. But I could see that the IR active wavenumbers in the output without any negative frequency. I am currently running version 6.8 of NWChem.

Below is my input:

echo
start molecule
title "RamanSolvation-CPCEstaggered"
charge 0
memory 5000 mb
geometry autosym 0.2
C -1.92943 0.97618 0.17175
C -0.38250 0.99560 0.15383
H -2.34114 0.23376 -0.54399
H -2.34138 1.98054 -0.06762
H -2.29646 0.69116 1.18224
C 0.19578 1.40831 -1.21741
C 0.09556 -0.45366 0.41732
C 0.05521 1.98483 1.25601
H -0.22556 -1.11834 -0.41475
H -0.34077 -0.87475 1.34693
H 1.20409 -0.50222 0.48216
C -0.60945 1.52340 -2.37793
C 0.53160 1.55418 2.51900
C 1.57748 1.67074 -1.34690
H -1.67056 1.32802 -2.36267
C -0.05103 1.87700 -3.60833
C 1.31331 2.13967 -3.70967
H -0.68120 1.94556 -4.48642
C 2.12605 2.04153 -2.58001
H 1.73928 2.41796 -4.66524
H 3.18460 2.25098 -2.66173
H 2.23172 1.60092 -0.48740
C 0.88474 2.48261 3.51972
C 0.80871 3.84763 3.24460
C 0.33694 4.29184 2.01101

C -0.05309 3.37371 1.03238
H 0.61426 0.50514 2.75383
H 0.25980 5.35392 1.81959
H -0.42859 3.75041 0.08989
O 1.26819 2.10182 4.76315
H 1.08716 4.56703 4.00489
C 0.89101 0.95970 5.32239
N 0.55847 -0.03835 5.80507
end

basis spherical
* library Def2-TZVP
end
basis "cd basis" spherical
* library "Weigend Coulomb Fitting"
end

dft
xc becke88 perdew86
disp vdw 3
noio
direct
grid xfine
convergence energy 1d-10
print low
iterations 1000
end
set quickguess t

driver
clear
maxiter 500

tight
end

set prop:newaoresp 1
set cphf:thresh 1.0d-8
set cphf:acc 1.0d-16
set cphf:maxiter 60

property
response 1 0.0570 # hartrees for 800 nm
damping 0.004
convergence 1e-6
end

raman
normal
lorentzian
width 10
end

task dft optimize
task dft raman

I really appreciate any help.

Regards,

Sophia

Edoardo Aprà

unread,

Feb 4, 2021, 2:39:02 PM2/4/21

to NWChem Forum

The CPHF numerical issue is likely to be due to the fact that your input file uses a "cd basis" for the optimization, while that task raman does not.

SInce the optimized geometries with and without "cd basis" are not going to be exactly the same, this might the cause for CPHF residual problem.

You can copy the optimized geometry from the output file, remove the "cd basis" and re-run to see if it makes any difference.

sophia sagala

unread,

Feb 12, 2021, 1:50:53 PM2/12/21

to nwchem...@googlegroups.com

Yes, removing the “cd basis” works and I also set the env. Variable PSM2_MEMORY=large.

But, the calculation did not reach completion and have the following error:

slurmstepd: error: Detected 1 oom-kill event(s) in StepId=157341.1 cgroup. Some of your processes may have been killed by the cgroup out-of-memory handler.

srun: error: cpu-2-36: task 6: Out Of Memory

I have also asked this issue with the system administrator of the system, and they have suggested some update on the script but still did not work and they are still working on it.

I wondered if anyone of you could help us give tips or a trick, that I can try from the input side or else ?

Here is the output of the calculation.

https://drive.google.com/file/d/1fRa-eMmUUlF8GhI-RlzOXuF9roLt3i4J/view?usp=sharing

Also, another question is how I can get the information about the depolarization ratio in the Raman calculation output? Does a certain parameter need to be added in the Raman block to get the depolarization ratio info or some information/results in the output that I should further use to calculate this ratio?

I really appreciate for any help and your time.

Regards,

Sophie

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/efe13c11-0afe-4070-a615-74ccf06975een%40googlegroups.com.

Edoardo Aprà

unread,

Feb 12, 2021, 3:21:39 PM2/12/21

to NWChem Forum

Adding the set fock:mirrmat f line before task raman might fix this Out of memory problem. However, it will make the calculation run slower.