[Wien] How to select the appropriate nband in hybrid functionals calculations

[Abo; Gavin Sky]

If the crash is due to not enough RAM, maybe it will show up similar to the exhaustion of RAM crash with nlvdw [1] if that is the case by checking for example with: grep hf /var/log/syslog

[1] http://zeus.theochem.tuwien.ac.at/pipermail/wien/2020-February/030339.html

On 5/15/2021 12:56 AM, Tran, Fabien wrote:

​Since your system is quite big, maybe the calculation crashed because of not enough RAM (hybrid calculations require more RAM than GGA).

Questions: Have you tried to run you calculations several times? How many cores/nodes are you using? Which kind of parallelization did you use (MPI or k-point)?


Not related to the crash: for the convergence criteria, use something like at least "-ec 0.001 -cc 0.001" to have reasonably converged SCF iterations.


________________________________
From: Wien <wien-bounces at zeus.theochem.tuwien.ac.at> on behalf of Yifan Ding <yfding0375 at foxmail.com>
Sent: Saturday, May 15, 2021 8:00 AM
To: wien
Subject: [Wien] How to select the appropriate nband in hybrid functionals calculations

Dear Prof. Fabien，

Thank you very much for reading and helping to solve my question.

I opened the file output.log, and there was only one line in it:

>   stop error

I found the output.log file for a previous successful example of hybrid functional computation, which shows the following:

ec cc and fc_conv 0 1 1
in cycle 2    ETEST: .9176672200000000   CTEST: .0349624
ec cc and fc_conv 0 1 1
in cycle 3    ETEST: .5190131650000000   CTEST: .0400105
ec cc and fc_conv 1 1 1

>   stop

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20210515/f62e61e7/attachment.htm>