[Wien] MPI error

[Peter Blaha]

As Laurence Marks mentioned, looking into all files you proposed, would 
cost a couple of hours. You have to check these files yourself and solve 
the problem or at least extract the most important information.


A few remarks:
>       > You need to link with the blacs library for openmpi.
>     I unsuccessfully recompiled wien2k by linking with the blacs library
>     for openmpias “mkl_blacs_openmpi_lp64” due to gfortran errors. 

The compile messages show that the serial version was compiled with 
ifort (and also the compiler options are for ifort). For mpi you used 
mpif90, however, this links to gfortran. Your compiler options for ifort 
are of course not compatible with this.
You need a different mpi-compiler, which links to ifort (maybe mpifort, 
this is what we have for our openmpi+ifort installation (with one i 
only!!) I would retry the installation but with mpifort as mpi-compiler)

>     The openmpi and fftw of the cluster are compiled with gfortran. So I
>     have also installed openmpi 4.1.0 and fftw3.3.9 in my home directory
>     after loading ifort and icc with the following commands:

Compilation looks ok, but:
your own openmpi is configured only for Gbit ethernet, not for 
Infiniband (please "read" the error message). If you really want to use 
your own openmpi, it needs to be configured for Infiniband.

In any case, a "good" supercomputer system should have "compatible" 
modules and you have to select the proper modules, which work together.

I'm pretty sure that everything is there what you need, but it looks 
that none of the sys-admins can really help or knows the details, 
how/which modules one should use and which commands one should use.

>     ./configure--prefix=/home/users/mollabashi/expands/openmpi CC=icc
>     F77=ifort FC=ifort--with-slurm --with-pmix --enable-shared
>     --with-hwloc=internal
>     ./configure--prefix=/home/users/mollabashi/expands/fftw MPICC=mpicc
>     CC=icc F77=ifort --enable-mpi --enable-openmp  --enable-shared
>     By this way, Wien2k compiled correctly as shown in the video in
>     _https://files.fm/u/rk3vfqv5g
>     <https://urldefense.com/v3/__https://files.fm/u/rk3vfqv5g__;!!Dq0X2DkFhyF93HkjWTBQKhk!AduEFeykF0ZeU4nMinZQqgGBUELDwJAo-Js9rRaimp_GI43yZzxiUL01UV7nLt3f-oLO0g$>_
>     link. But mpi run does not run due to the error about openmpi as in:
>     _https://files.fm/u/tcz2fvwpg
>     <https://urldefense.com/v3/__https://files.fm/u/tcz2fvwpg__;!!Dq0X2DkFhyF93HkjWTBQKhk!AduEFeykF0ZeU4nMinZQqgGBUELDwJAo-Js9rRaimp_GI43yZzxiUL01UV7nLt107mI0Og$>_.
>     The .bashrc, submit.sh and slurm.out filesare in
>     _https://files.fm/u/dnrrwqguy
>     <https://urldefense.com/v3/__https://files.fm/u/dnrrwqguy__;!!Dq0X2DkFhyF93HkjWTBQKhk!AduEFeykF0ZeU4nMinZQqgGBUELDwJAo-Js9rRaimp_GI43yZzxiUL01UV7nLt0vfsNB5A$>_
>     link.
>     Would you please guide me how to solve the gfortran errors?
>     Should I install openmpi with another configuration to solve the
>     slurm error of mpi calculation?
>     Sincerely yours,
>     Leila
> 
> 
> -- 
> Professor Laurence Marks
> Department of Materials Science and Engineering
> Northwestern University
> www.numis.northwestern.edu <http://www.numis.northwestern.edu/>
> "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
> 

-- 
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300             FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.at    WIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at
-------------------------------------------------------------------------