[Wien] Elpa 2025

[Laurence Marks]

It compiled fine. However the MPI+omp does not work, it only does MPI when
you check the Wall/CPU times.

___
Emeritus Professor Laurence Marks (Laurie)
Department of Materials Science and Engineering, Northwestern University
www.numis.northwestern.edu
"Research is to see what everybody else has seen, and to think what nobody
else has thought" Albert Szent-Györgyi

On Sat, Apr 26, 2025, 03:03 Fecher, Gerhard <fecher at uni-mainz.de> wrote:

> Dear Laurence
> I compiled ELPA 2025.01.001 without problems using the OneAPI 2025.1.0
>
> what is your problem ?
> is it ifort or icc ? too old compilers ?
> (Note my programms run partially much faster when using the latest ifx/icx
> compared to older 2021 or 22 OneAPI versions,
> therefore I did not install the old stuff on new computers anymore)
>
> If I use --enable-omp without specifying -qopenmp in the configuration,
> the the icx compiler complains that one should use -qopenmp instead of
> -fopenmp.
> Obviously the configure script does not distinguish wether one uses icx or
> a different compiler (in old icc versions it was -openmp instead of
> -qopenmp !).
> However, none of the ELPA checks fails.
>
> The complains (warnings) of the icx about -qopenmp vanish in the following
> configuration,
> where the checks also run without problems:
> ./configure --prefix=$HOME/lib/elpa-intel \
> CC=mpiicx CXX=mpiicpx CFLAGS="-O3 -xCORE-AVX512 -qopenmp"  \
> FC=mpiifx FCFLAGS="-O3 -xCORE-AVX512 -fp-model=precise -qopenmp" \
> SCALAPACK_FCFLAGS="-I${MKLROOT}/include" \
> SCALAPACK_LDFLAGS="-L${MKLROOT}/lib -lmkl_scalapack_lp64 \
> -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -lmkl_blacs_intelmpi_lp64
> -liomp5 -lpthread -lm -ldl \
> -Wl,-rpath,${MKLROOT}/lib" \
> --enable-option-checking=fatal --with-mpi=yes --enable-openmp \
> --disable-c-tests --disable-cpp-tests \
> --enable-avx2 --enable-avx512-kernels
>
> My problem is how to make the fastest ELPA library on systems with AMD
> EPYC processors (usually -xCORE... fails on AMD's),
> the configure script for icx does not recognice the processor and that it
> has the AVX512 extensions
> indeed, Intel is not interested in support of AMD processors,
> does someone know about the GNU support for AVX512 on AMD's, is it well
> implemented in gcc ?
>
> Ciao
> Gerhard
>
> DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy:
> "I think the problem, to be quite honest with you,
> is that you have never actually known what the question is."
>
> ====================================
> Dr. Gerhard H. Fecher
> Institut of Physics
> Johannes Gutenberg - University
> 55099 Mainz
> ________________________________________
> Von: Wien [wien-bounces at zeus.theochem.tuwien.ac.at] im Auftrag von
> Laurence Marks [laurence.marks at gmail.com]
> Gesendet: Freitag, 25. April 2025 22:59
> An: A Mailing list for WIEN2k users
> Betreff: [Wien] Elpa 2025
>
> Has anyone got the 2025 version of elpa running in hybrid mode
> (--enable-openmp and -qopenmp) with ifort/icc? 2024.05.001 seems to be
> fine, but with 2025.01.001 the omp does not work for me in hybrid mode.
> (ifort 2021.1.1 as I know it is stable).
>
>
> --
> Emeritus Professor Laurence Marks (Laurie)
> Northwestern University
> Webpage<http://www.numis.northwestern.edu> and Google Scholar link<
> http://scholar.google.com/citations?user=zmHhI9gAAAAJ&hl=en>
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought", Albert Szent-Györgyi
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20250426/0be52da8/attachment.htm>