Page 1 of 1
Problem in computing the exchange part of the self-energy.
Posted: Thu Nov 17, 2022 2:16 pm
by Bikash
Dear Developers,
I am trying to calculate the quasiparticle band structure of my system.
But in the calculation of the HF self-energy part, my job was automatically killed
without writing the output file. I have attached all the necessary files.
Is this a memory issue or any other problem?
with regards,
Bikash
TIFR, Mumbai, India
Re: Problem in computing the exchange part of the self-energy.
Posted: Thu Nov 17, 2022 2:33 pm
by Daniele Varsano
Dear Bikash,
It seems it is crashing when computing the electronic density and it's not easy to spot the problem. I do not think it is a memory problem.
I suggest you the following test:
*remove all the ndb.* databases in the SAVE directory
*rerun a setup without assigning MaxGvecs variable.
*rerun the HF calculation.
If it does not work, you can try to download the latest release of the code (5.1) and compile by adding the flag --enable-int-linalg and retry.
Best,
Daniele
Re: Problem in computing the exchange part of the self-energy.
Posted: Mon Nov 21, 2022 6:45 am
by Bikash
Dear Sir,
The problem remains the same even if I use the latest version
compiled using the flag suggested by you.
What I find is that if I reduce the MaxGvecs variable to 10 Ry,
It is working fine.
Could you please tell me how I rerun a setup without assigning MaxGvecs variable?
with regards,
Bikash.
Re: Problem in computing the exchange part of the self-energy.
Posted: Wed Nov 23, 2022 8:14 am
by Daniele Varsano
Dear Bikash,
10Ry as MaxGvecs sounds a but small.
In order to not fix the MaxGvecs, delete your ndb.* files in the SAVE directory and just run yambo without building any input file.
Best,
Daniele