[Q-e-developers] Using QE as a benchmark code for machine procurement

[Nicola Marzari]

Dear Zhengji Zhao,

wonderful news, indeed. I believe there was a large test case
(graphene on Ir) that had been submitted to PRACE - it would be
good to see if Stefano dG or others have it.

Heather Kulik (cc'd here) had some experience on runnning pwscf on
some of the livermore machines, so maybe she could comment on what could 
be a suitable large-scale case - or maybe someone else in the
quantum simulations group there.

It would be great to have the large test cases also for cp (also part of 
Q-E) - that is a very good code for large-scale simulations.

If you'd like, we have standard inputs for ab-initio md of ferroelectric
perovskites, 320 atoms in the unit cell - this is a good medium size system.

All the best,

			nicola



On 1/19/11 8:19 PM, Zhengji Zhao wrote:
> Dear Developers,
>
> We are considering to use Quantum Espresso code as one of the benchmark
> codes for our next machine procurement, as we have more and more
> materials science and chemistry users use this code for their research.
> I am writing to get some advice from you (developers). I appreciate if
> you can help me with the following questions:
>
> 1) Do you know any other super computing centers (and other parties)
> that use QE as a benchmark code for their HPC computer procurement? If
> yes, we are interested in getting their benchmark suite (makefiles, test
> cases, reference runs, etc.) for reference. Any information towards this
> would be highly appreciated.
>
> 2) If there is no established benchmark suite for the HPC computer
> procurement, we would like to work out such a benchmark suite. Could you
> give us some guidance when preparing such a benchmark suite? We want to
> come up  with 3 test cases (input files) that can use a small (eg., 32
> cores), medium(eg., 256 cores), and large (eg., ~1024 cores and more if
> possible) number of processor cores. Any help towards this, such as,
> suitable input files or info about it,  comments/info about the parallel
> scaling of the codes, ..., would be highly appreciated as well.
>
> Regards,
>
> Zhengji Zhao
> NERSC User Services
> Lawrence Berkeley National Laboratory
> _______________________________________________
> Q-e-developers mailing list
> Q-e-developers at qe-forge.org
> http://qe-forge.org/mailman/listinfo/q-e-developers


-- 
----------------------------------------------------------------------
Prof Nicola Marzari    Department of Materials    University of Oxford
Chair of Materials Modelling  Director, Materials Modelling Laboratory
nicola.marzari at materials.ox.ac.uk     http://mml.materials.ox.ac.uk/NM