CP2K 8.1: Runtime error running on more than one node

[Mauro Sgroi]

Dear Developers,

I'm writing to ask you help for setting CP2K on a HPC server.

Our IT compiled the code using gcc-8.3.0, openmpi 4.1.0 and the toolchain:

cmake-3.18.5 cosma-2.2.0 elpa-2020.05.001 fftw-3.3.8 gsl-2.6 hdf5-1.12.0 libint-v2.6.0-cp2k-lmax-5 libvdwxc-0.4.0 libvori-201229 libxc-4.3.4 libxsmm-1.16.1 openblas-0.3.10 scalapack-2.1.0 sirius-7.0.0 SpFFT-0.9.13 spglib-1.16.0 

The obtained code works fine on a single node of our server. When we launch it with the command:

mpiexec --prefix ${MPI_INST_DIR} -x PATH -x LD_LIBRARY_PATH -n $CORES -machinefile machine_file ${INST_DIR}/cp2k-8.1/exe/local/cp2k.${SOLVER} -inp ${INPUT} 2>&1  

on more than one node we get the error:

[btl_openib_component.c:1699:init_one_device] error obtaining device attributes for mlx5_0 errno says Protocol not supported  

Could you please give us some advice?

Thanks a lot in advance and best regards,

Mauro Sgroi.

Centro Ricerche FIAT.

Italy.

[Alin Marin Elena]

Hi Mauro,

seen this before with other codes... the issue invariable ended up due
to openmpi not playing nicely with the queuing system.

if you want to see if a simple MPI hello world using the above mpi
actually works....

Regards,
Alin
Without Questions there are no Answers!
______________________________________________________________________
Dr. Alin Marin ELENA
http://alin.elena.space/
______________________________________________________________________

> --
> You received this message because you are subscribed to the Google Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/cc9bb9af-4e72-4667-ad33-ef04edf7a70bn%40googlegroups.com.

[Mauro Sgroi]

Dear Alin,

thanks a lot for the reply.

With those mpi we are running smoothly Quantum Espresso and other codes.

So it seems strange to find this error.

Best regards,

Mauro Sgroi.

[Tiziano Müller]

Hi Mauro,

there are some things you can try:

* build a CP2K without COSMA, ELPA, SIRIUS, SpFFT and see whether you
can reproduce it with that (simply create an arch-file as a copy of the
toolchain-generated ones with the respective -D..., -l..., -I... flags
removed)
* check the toolchain build logs of ELPA and COSMA (cmake.log/make.log)
to verify that the correct MPI implementation was picked up during
compilation, see also [1]

My guess since other (less complex) software works is that either there
is/was a mismatch of OpenMPI implementations and/or that one of the
dependencies tries to do some advanced MPI ops (RMA, I/O) which is not
configured on your cluster (CP2K has RMA off by default).

Best,
Tiziano

[1] https://github.com/cp2k/cp2k/issues/1351

On 2/15/21 10:07 PM, Mauro Sgroi wrote:
> Dear Alin,
> thanks a lot for the reply.
> With those mpi we are running smoothly Quantum Espresso and other codes.
> So it seems strange to find this error.
> Best regards,
> Mauro Sgroi.
> Il giorno lunedì 15 febbraio 2021 alle 18:43:31 UTC+1 alinm...@gmail.com
> ha scritto:
>
> Hi Mauro,
>
> seen this before with other codes... the issue invariable ended up due
> to openmpi not playing nicely with the queuing system.
>
> if you want to see if a simple MPI hello world using the above mpi
> actually works....
>
> Regards,
> Alin
> Without Questions there are no Answers!
> ______________________________________________________________________
> Dr. Alin Marin ELENA

> http://alin.elena.space/ <http://alin.elena.space/>

> <https://groups.google.com/d/msgid/cp2k/cc9bb9af-4e72-4667-ad33-ef04edf7a70bn%40googlegroups.com>.

>
>
> --
> You received this message because you are subscribed to the Google
> Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to cp2k+uns...@googlegroups.com

> <mailto:cp2k+uns...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/cp2k/7b611401-f960-464c-9394-0b3add2aad6dn%40googlegroups.com
> <https://groups.google.com/d/msgid/cp2k/7b611401-f960-464c-9394-0b3add2aad6dn%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Tiziano Müller
University of Zurich
Department of Chemistry
Winterthurerstrasse 190
CH-8057 Zürich

Tel: +41 44 63 54234
www.chem.uzh.ch
tiziano...@chem.uzh.ch

[Mauro Sgroi]

Dear Tiziano,

thanks a lot.

We solved changing the compilation flags of COSMA and ELPA.

Best regards,

Mauro.