Hi all,<br />I'm trying to build cuda_ssmp version of CP2K 2024.1, without the MPI support (not really necessary since the machine has only one GPU card?).<br />The build process goes well, but the regtest failed as:<br /><br /><font face="Courier New">[root@gpu01 cp2k-2024.1]# make ARCH=local_cuda VERSION=ssmp test<br />(......)<br />========= Python (ssmp) =========<br />/usr/bin/env python3 --version<br />Python 3.6.8<br />----------------------- External Modules ---------------------------------<br />DBCSR Version: 2.6.0 (2023-07-10)<br />---------------------------- Modules -------------------------------------<br />Currently Loaded Modulefiles:<br /> 1) gcc-9.5.0/gcc 2) cuda/11.7.1 3) tools/cmake-3.28.1 4) tools/git-2.43.0<br />*************************** Testing started ****************************<br />ERROR: cuInit failed with error: 999 /cluster/bld/cp2k-2024.1/src/offload/offload_library.c 57<br />Program received signal SIGABRT: Process abort signal.<br />Backtrace for this error:<br />#0 0x7f8837d243ff in ???<br />#1 0x7f8837d24387 in ???<br />#2 0x7f8837d25a77 in ???<br />#3 0x30fe53b in offload_init<br /> at /cluster/bld/cp2k-2024.1/src/offload/offload_library.c:58<br />#4 0xaf6e0b in __f77_interface_MOD_init_cp2k<br /> at /cluster/bld/cp2k-2024.1/src/f77_interface.F:234<br />#5 0x455f88 in cp2k<br /> at /cluster/bld/cp2k-2024.1/src/start/cp2k.F:284<br />#6 0x40ebdc in main<br /> at /cluster/bld/cp2k-2024.1/src/start/cp2k.F:44<br />Could not parse feature flags.</font><br /><br />The machine has a RTX A6000 card, and CUDA toolkit 12.3.2 was used:<br /><font face="Courier New">+---------------------------------------------------------------------------------------+<br />| NVIDIA-SMI 545.23.08 Driver Version: 545.23.08 CUDA Version: 12.3 |<br />|-----------------------------------------+----------------------+----------------------+<br />| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |<br />| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |<br />| | | MIG M. |<br />|=========================================+======================+======================|<br />| 0 NVIDIA RTX A6000 Off | 00000000:3B:00.0 Off | Off |<br />| 30% 36C P0 67W / 300W | 2MiB / 49140MiB | 2% Default |<br />| | | N/A |<br />+-----------------------------------------+----------------------+</font><br /><br />The OS is CentOS 7.9, and I used a source-built GCC 9.5.0 as loadable module.<br />The toolchain was built with:<br /><br /><font face="Courier New">./install_cp2k_toolchain.sh --mpi-mode=no --with-cmake=system -j 16 --enable-cuda=yes --gpu-ver=A100</font><br /><br />and then in the CP2K source folder:<br /><br /><font face="Courier New">make -j 16 ARCH=local_cuda VERSION=ssmp</font><br /><br />However, the CPU ssmp version (ARCH=local) was built and can pass all the regtests successfully.<br />Any suggestions on making the CUDA version works?<br /><br />Mike<br />
<p></p>
-- <br />
You received this message because you are subscribed to the Google Groups "cp2k" group.<br />
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="mailto:cp2k+unsubscribe@googlegroups.com">cp2k+unsubscribe@googlegroups.com</a>.<br />
To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/cp2k/affc7786-0b86-4e51-8d40-22247d0d2ab9n%40googlegroups.com?utm_medium=email&utm_source=footer">https://groups.google.com/d/msgid/cp2k/affc7786-0b86-4e51-8d40-22247d0d2ab9n%40googlegroups.com</a>.<br />