[QE-users] GRIR443 benchmark - is it running properly?

Steve Brasier steveb at stackhpc.com
Fri Dec 6 14:42:30 CET 2019


Thank you Paolo that looks much better now!

---

It turns out it doesn't ignore the -nimage option, but interestingly both
with and without it the header still reported the same number of
processors. However the memory usage and communication now look more
reasonable.

Does anyone know where the command-line options for pw.x are listed?
Neither the general nor the PW user guide seem to. The general user guide
describes parallelisation (
https://www.quantum-espresso.org/Doc/user_guide/node18.html) but not which
options PW accepts.

thank you
Steve Brasier

On Fri, 6 Dec 2019 at 12:15, Paolo Giannozzi <p.giannozzi at gmail.com> wrote:

> On Fri, Dec 6, 2019 at 11:50 AM Steve Brasier <steveb at stackhpc.com> wrote:
>
> srun --mpi=pmix /q-e-qe-6.4.1/bin/pw.x -nimage 4 -in grir443.in
>> -disk_io='none'
>>
>
> -nimage is not used. Command-line option -disk_io is nonexistent.
>
> there's zero communications between the 4 nodes I'm using. Is that
>> expected behaviour for pw.x, or does that suggest something is wrong?
>>
>
> I think pw.x ignores option -nimage 4, but if it doesn't, you are running
> 4 copies of the same run on a single processor. In the header there is a
> message saying how many processors the code is running on.
>
> Paolo
>
>
>
>> Thanks for any help, or even pointers to GRIR443 tutorials or results.
>>
>> regards
>> Steve Brasier
>>
>>
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>
>
>
> --
> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20191206/91ca6d70/attachment.html>


More information about the users mailing list