[QE-users] Optimizing QE Install

[Paolo Giannozzi]

The performances of a large and complex code are affected by a number of
factors: hardware[1], compilation [2], mathematical libraries[3],
parallelism[4], obvious[5] and less obvious[6] conflicts for resources,
phase of the moon[7] and alignment of planets[8]...

Paolo

[1] Not only CPU speed but also speed of memory access, size of cache,...As
a rule: the cheaper the hardware, the slower the speed
[2] Can make a sizable difference, but, as Lorenzo says, not a factor 10
[3] These, especially FFTs and linear algebra, can also make a sizable
difference
[4] There are many ways to use a given number of processors; just
increasing their number is not always the right thing to do
[5] Some other process running on the same processor
[6] Memory access conflicts, unwanted OpenMP threads, ...
[7-8] By far the most important factors:-) Many years ago, I watched in
disbelief the same run taking 27 s in one case, 30 in another,
reproducibly. The difference? a single print statement, executed once.


On Sat, Jan 23, 2021 at 10:22 PM Robert Stanton <stantor at clarkson.edu>
wrote:

> Dear all,
>
>      I am able to successfully install parallel QE and get everything to
> run smoothly, however after testing against an install on an externally
> maintained cluster I'm finding my speed to be slower by a factor of 5-10x
> when keeping the system, core number, QE version, etc. the same.
>
>     I'm not sure the degree to which this could be a hardware discrepancy
> (Xeon Silvers vs new EPYCs), v.s. how much could be a sub-optimal
> compilation of QE. The slower QE install is being done with blas/lapack
> 3.8.0 and openmpi 1.10.7. I'm just curious if there is anything very
> clearly not ideal about this environment, or if it is likely something I'll
> just need to troubleshoot with smaller settings? Thanks so much in advance
> for any insights you can provide!
>
> Regards,
> Robert Stanton
> Graduate Student
> Clarkson University
>
> _______________________________________________
> 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/20210124/09f55b38/attachment.html>