Errors with COHSEX

Concerns issues with computing quasiparticle corrections to the DFT eigenvalues - i.e., the self-energy within the GW approximation (-g n), or considering the Hartree-Fock exchange only (-x)

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano

Post Reply
Quxiao
Posts: 96
Joined: Fri Mar 26, 2021 11:27 am

Errors with COHSEX

Post by Quxiao » Thu Sep 15, 2022 5:05 am

Dear all,
I want to do a GoWo and BSE calculation on top of COHSEX . First, I do a COHSEX calculation by using yambo_sc "yambo_sc -v h+coh+sex -X s -s" (viewtopic.php?p=10031&hilit=COHSEX+level+GW#p10031). But it stopped at the following step
[06] Self-Consistent COH+SEX+HARTREE
====================================


[SC] Self-Consistency (SC) : yes
[SC] Restart (quick rebuild) : no
[SC] Restart + SC : no
[SC] Non SC single shot : no
[SC] Bands : 1 156
[SC] Bands mixing : 100.0000
[SC] Use non-local mixing : no
[SC] Preconditioning : simple
[SC] Diagonal approximation : no
[SC] N (N-1) order mixing : 0.500000
[SC] Imposed accuracy on density : 0.100000E-4
[SC] on magnetization: 0.100000E-4
[SC] on energies : 0.010000 [eV]


[WARNING] HXC collisions not found. Potential is computed on the fly.

Here I post the input and log file. Would you please help me check which variables is not correct? Thank you!
You do not have the required permissions to view the files attached to this post.
Quxiao in BIT,calculate the exciton

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: Errors with COHSEX

Post by Daniele Varsano » Thu Sep 15, 2022 7:54 am

Dear Quxiao,

the reported Warning is harmless. The problem here seems to be related to a memory issue:

Code: Select all

[MEMORY] Alloc WF%c( 18.47040 [Gb]) TOTAL:  18.61355 [Gb] (traced)  1.849260 [Gb] 
Please note that the yambo_sc release is experimental and your calculation seems to be quite demanding.

You can try to run the calculation using more CPUs (parallelizing on bands only):

Code: Select all

SE_CPU= "1 1 Ncpu"                       # [PARALLEL] CPUs for each role
SE_ROLEs= "q qp b"     


or reduce the number of bands to be calculated.
Best,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Post Reply