[CP2K-user] [CP2K:20407] ELPA ERROR with cp2k container

Ivan Bosch ivan.bosch at irbbarcelona.org
Wed Jul 3 07:22:53 UTC 2024


Yes, I've checked it. 

I'm attaching the output files I have obtained from the tests for the 
container (master version) and for the spack module (v2024.1) that I 
mentioned in the previous message so you can have all the information.

Thanks

On Tuesday, July 2, 2024 at 4:03:30 PM UTC+2 Krack Matthias wrote:

> Thanks for your report.
>
> Did you also check with the print key which ELPA kernel is actually used?
>
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Ivan 
> Bosch <ivan.... at irbbarcelona.org>
> *Date: *Tuesday, 2 July 2024 at 14:36
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *Re: [CP2K:20400] ELPA ERROR with cp2k container
>
> Thanks for your response,
>
> I ran a test with the container using ELPA, which is when the warning in 
> question appears, and another with ScalaPack with an input for only 5 
> steps. These are the results I obtained:
>
> [root at irblogin01 outputs_tests]# grep diag devel_*elpa*_autokernel_test.out 
> | tail -4 
> qs_ot_get_derivative_diag 160 12.0 0.002 0.003 3.273 4.061 
> qs_ot_p2m_diag 161 11.0 0.026 0.104 2.773 2.863 
> cp_fm_diag_elpa 161 13.0 0.000 0.001 2.013 2.015 
> cp_fm_diag_elpa_base 161 14.0 0.933 1.876 0.941 1.889
> [root at irblogin01 outputs_tests]# grep diag devel_*elpa*_avx512_block6kernel_test.out 
> | tail -4 
> qs_ot_get_derivative_diag 160 12.0 0.002 0.003 3.251 3.954 
> qs_ot_p2m_diag 161 11.0 0.026 0.103 2.772 2.862 
> cp_fm_diag_elpa 161 13.0 0.000 0.001 2.011 2.012 
> cp_fm_diag_elpa_base 161 14.0 0.933 1.874 0.940 1.888 
> [root at irblogin01 outputs_tests]# grep diag devel_*scalapack*_test.out | 
> tail -4 
> qs_ot_get_derivative_diag 160 12.0 0.002 0.003 3.319 4.190 
> qs_ot_p2m_diag 161 11.0 0.026 0.101 2.090 2.175
>
> In this case, there is no significant acceleration when using ELPA. 
>
> However, with version 2024.1 installed with Spack, where I have verified 
> that ELPA works well, I get these results:
>
> [root at irblogin01 outputs_tests]# grep diag 2024.1_*scalapack*_test.out | 
> tail -4 
> qs_ot_p2m_diag 161 11.0 0.010 0.017 1.828 1.841 
> qs_ot_get_derivative_diag 160 12.0 0.002 0.002 1.410 1.421 
> [root at irblogin01 outputs_tests]# grep diag 2024.1_*elpa*_avx512_block6kernel_test.out 
> | tail -4 
> qs_ot_get_derivative_diag 160 12.0 0.002 0.002 1.397 1.408 
> qs_ot_p2m_diag 161 11.0 0.010 0.017 1.190 1.205 
> cp_fm_diag_elpa 161 13.0 0.000 0.001 0.816 0.817 
> cp_fm_diag_elpa_base 161 14.0 0.321 0.646 0.327 0.662
>
> In these, I do observe an improvement in the timings.
>
>  
>
> On Tuesday, July 2, 2024 at 10:56:58 AM UTC+2 Krack Matthias wrote:
>
> You can use the print key 
> https://manual.cp2k.org/trunk/CP2K_INPUT/GLOBAL/PRINT_ELPA.html to get 
> more information about the ELPA usage.
>
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Ivan 
> Bosch <ivan.... at irbbarcelona.org>
> *Date: *Tuesday, 2 July 2024 at 10:43
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *Re: [CP2K:20398] ELPA ERROR with cp2k container
>
> Sure, but does this warning mean that the ELPA library is not being used 
> to diagonalize?
>
> Thanks
>
> On Tuesday, July 2, 2024 at 10:38:39 AM UTC+2 Krack Matthias wrote:
>
> Hi
>
>  
>
> That’s just a warning which is not specific to the use of containerized 
> CP2K and/or singularity. I usually ignore it.
>
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Ivan 
> Bosch <ivan.... at irbbarcelona.org>
> *Date: *Tuesday, 2 July 2024 at 09:43
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *[CP2K:20394] ELPA ERROR with cp2k container
>
> Hi, I've made a build of the cp2k-container provided by 
> https://github.com/cp2k/cp2k-containers, specifically using 
> master_openmpi_skylake-avx512_psmp.Dockerfile, and it gives the following 
> error regarding ELPA:
>
> bash
>
> Copy code
>
> *** WARNING in fm/cp_fm_elpa.F:522 :: Setting real_kernel for ELPA failed 
> ***. 
>
> This has happened with different cp2k containers, not just this one, even 
> when pulling directly from Docker Hub. I am using SingularityCE v4.0, and 
> this is the output from within the container when running cp2k -v:
>
> bash
>
> Copy code
>
> CP2K version 2024.1 (Development Version) Source code revision git:5b092d7 
> cp2kflags: omp libint fftw3 libxc libgrpp pexsi elpa parallel mpi_f08 
> scalapack cosma quip deepmd xsmm plumed2 spglib libdftd4 sirius libvori 
> libbqb libtorch libvdwxc hdf5 compiler: GCC version 11.4.0 compiler 
> options: -cpp -I /opt/cp2k/tools/toolchain/install/openmpi-4.1.6/include -I 
> /opt/cp2k/tools/toolchain/install/openblas-0.3.27/include -I 
> /opt/cp2k/tools/toolchain/install/fftw-3.3.10/include -I 
> /opt/cp2k/tools/toolchain/install/libint-v2.6.0-cp2k-lmax-5/include -I 
> /opt/cp2k/tools/toolchain/install/libxc-6.2.2/include -I 
> /opt/cp2k/tools/toolchain/install/libgrpp-main-20231215/include -I 
> /opt/cp2k/tools/toolchain/install/libxsmm-1.17/include -I 
> /opt/cp2k/tools/toolchain/install/COSMA-2.6.6/include -I 
> /opt/cp2k/tools/toolchain/install/elpa-2024.03.001/cpu/include/elpa_openmp-2024.03.001/modules 
> -I 
> /opt/cp2k/tools/toolchain/install/elpa-2024.03.001/cpu/include/elpa_openmp-2024.03.001/elpa 
> -I /opt/cp2k/tools/toolchain/install/scotch-6.0.0/include -I 
> /opt/cp2k/tools/toolchain/install/superlu_dist-6.1.0/include -I 
> /opt/cp2k/tools/toolchain/install/pexsi-1.2.0/include -I 
> /opt/cp2k/tools/toolchain/install/quip-0.9.10/include -I 
> /opt/cp2k/tools/toolchain/install/gsl-2.7/include -I 
> /opt/cp2k/tools/toolchain/install/libdeepmd_c-2.2.7/include -I 
> /opt/cp2k/tools/toolchain/install/hdf5-1.14.2/include -I 
> /opt/cp2k/tools/toolchain/install/libvdwxc-0.4.0/include -I 
> /opt/cp2k/tools/toolchain/install/spglib-2.3.1/include -I 
> /opt/cp2k/tools/toolchain/install/SpFFT-1.1.0/include -I 
> /opt/cp2k/tools/toolchain/install/SpLA-1.6.1/include/spla -I 
> /opt/cp2k/tools/toolchain/install/sirius-7.5.2/include -I 
> /opt/cp2k/tools/toolchain/install/dftd4-3.6.0/include -I 
> /opt/cp2k/tools/toolchain/install/dftd4-3.6.0/include/dftd4/GNU-11.4.0 -I 
> /opt/cp2k/tools/toolchain/install/dftd4-3.6.0/include/mctc-lib/GNU-11.4.0 
> -I /opt/cp2k/src/start/ -I /opt/cp2k/obj/local/psmp/exts/dbcsr -I 
> /opt/cp2k/tools/toolchain/install/openmpi-4.1.6/include -I 
> /opt/cp2k/tools/toolchain/install/openmpi-4.1.6/lib -imultiarch 
> x86_64-linux-gnu -D_REENTRANT -D __LIBXSMM -D __parallel -D __MPI_F08 -D 
> __FFTW3 -D __LIBINT -D __LIBXC -D __LIBGRPP -D __SCALAPACK -D __COSMA -D 
> __ELPA -D __LIBPEXSI -D __QUIP -D __GSL -D __PLUMED2 -D __DEEPMD -D __HDF5 
> -D __LIBVDWXC -D __SPGLIB -D __LIBVORI -D __LIBTORCH -D __SPFFT -D __SPLA 
> -D __SIRIUS -D __DFTD4 -D __COMPILE_ARCH="local" -D __COMPILE_DATE="Mon Jul 
> 1 20:22:13 UTC 2024" -D __COMPILE_HOST="5bc8c4d74319" -D 
> __COMPILE_REVISION="git:5b092d7" -D __DATA_DIR="/opt/cp2k/data" -D 
> __SHORT_FILE__="start/cp2k.F" -mtune=skylake-avx512 -march=x86-64 -g -O3 
> -Werror=aliasing -Werror=ampersand -Werror=c-binding-type 
> -Werror=intrinsic-shadow -Werror=intrinsics-std -Werror=line-truncation 
> -Werror=tabs -Werror=target-lifetime -Werror=underflow 
> -Werror=unused-but-set-variable -Werror=unused-variable 
> -Werror=unused-dummy-argument -Werror=unused-parameter -Werror=unused-label 
> -Werror=conversion -Werror=zerotrip -Wno-maybe-uninitialized 
> -Wuninitialized -Wuse-without-only -std=f2008 -fno-omit-frame-pointer 
> -fopenmp -funroll-loops -fbacktrace -ffree-form -fimplicit-none 
> -fallow-argument-mismatch 
> -fpre-include=/usr/include/finclude/math-vector-fortran.h 
>
> Could you help me troubleshoot and resolve this issue?
>
> -- 
> 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... at googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/e07c8a89-84d7-49a7-8911-faada06530d1n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/e07c8a89-84d7-49a7-8911-faada06530d1n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> -- 
> 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... at googlegroups.com.
>
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/73bc7691-1ad9-493e-9641-6b5fd631ffa1n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/73bc7691-1ad9-493e-9641-6b5fd631ffa1n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> -- 
> 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... at googlegroups.com.
>
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/0135c2f3-5810-4ca8-ad53-eb3aa1a40839n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/0135c2f3-5810-4ca8-ad53-eb3aa1a40839n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>

-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/4831d24c-bb86-4b44-9355-f9f647295702n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240703/ec253327/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_cp2k.zip
Type: application/x-zip
Size: 110433 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240703/ec253327/attachment-0001.bin>


More information about the CP2K-user mailing list