[QE-users] Case study: KPOINT and band parallelization in Quantum ESPRESSO

[Paolo Giannozzi]

On Fri, Dec 17, 2021 at 6:42 AM pranav kumar <prnvkmr4 at gmail.com> wrote:

*nk nb time(sec)*
>
> 1 1 2062 1 1633 1 1444 1 1386 1 1388 1 13112 1 141
>
> this is the expected behavior: you are decreasing the number of processors
for plane-wave parallelization while increasing the number of processors
for k-point parallelization. Given the size of the system (not too big) and
the number of k-points (quite a lot) it is convenient to exploit also
k-point parallelization.

The "band" parallelization should not be used as a replacement of
plane-wave or k-point parallelization but only in addition to, for some
specific cases (e.g. hybrid functionals)

Paolo

> In second run, I fixed nk at 3 and varied nb and Here is the result
>
> *nk nb time(sec)* 3  1 1443  2 1813  4 2553  8 426
>
> Why is run-time increasing with increase in nb in second run? I was
> hoping for decreasing in trend.
>
>
> *Note: QE v6. 8 is compiled with intel-oneapi and internal FFT library*
>
> --
>
> PRANAV KUMAR
>  Research Scholar
> IIT MADRAS (AM18D011)
>
>
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> 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 206, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20211217/5e1314a9/attachment.html>