<div dir="ltr">Hi,<div>According to the CUDA documents I found, the driver CUDA support version and installed (or loaded) CUDA runtime version can be different.</div><div>It should work as long as the CUDA runtime version <= driver CUDA support version for backward compatibility.</div><div>And again, the error I got from CUDA runtime 12.2 and 11.7 are identical.</div><div>I'm trying to change the --gpu-ver values and see if if works.</div><div><br></div><div>Mike</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">Krack Matthias <<a href="mailto:matthias.krack@psi.ch">matthias.krack@psi.ch</a>> 於 2024年2月26日 週一 下午9:06寫道:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div class="msg4048358605615673235">





<div lang="en-CH" style="overflow-wrap: break-word;">
<div class="m_4048358605615673235WordSection1">
<p class="MsoNormal"><span lang="DE-CH" style="font-size:11pt">Hi
<u></u><u></u></span></p>
<p class="MsoNormal"><span lang="DE-CH" style="font-size:11pt"><u></u> <u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt">This looks like an issue with the CUDA installation. You load the module cuda/11.7.1, whereas nvidia-smi returns “CUDA Version 12.2”. Maybe, that’s something to check.<u></u><u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt"><u></u> <u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt">Best<u></u><u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt"><u></u> <u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt">Matthias<u></u><u></u></span></p>
<p class="MsoNormal"><span lang="EN-US" style="font-size:11pt"><u></u> <u></u></span></p>
<div id="m_4048358605615673235mail-editor-reference-message-container">
<div>
<div style="border-right:none;border-bottom:none;border-left:none;border-top:1pt solid rgb(181,196,223);padding:3pt 0cm 0cm">
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:12pt;margin-left:36pt">
<b><span style="font-size:12pt;font-family:Aptos,sans-serif;color:black">From:
</span></b><span style="font-size:12pt;font-family:Aptos,sans-serif;color:black"><a href="mailto:cp2k@googlegroups.com" target="_blank">cp2k@googlegroups.com</a> <<a href="mailto:cp2k@googlegroups.com" target="_blank">cp2k@googlegroups.com</a>> on behalf of Mike Chen <<a href="mailto:mike.scchen@gmail.com" target="_blank">mike.scchen@gmail.com</a>><br>
<b>Date: </b>Sunday, 25 February 2024 at 14:03<br>
<b>To: </b>cp2k <<a href="mailto:cp2k@googlegroups.com" target="_blank">cp2k@googlegroups.com</a>><br>
<b>Subject: </b>[CP2K:19967] Built cuda.ssmp version successfully but failed on regtest<u></u><u></u></span></p>
</div>
<p class="MsoNormal" style="margin-left:36pt"><span style="font-size:11pt">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>
</span><span style="font-size:11pt;font-family:"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.</span><span style="font-size:11pt"><br>
<br>
The machine has a RTX A6000 card, and CUDA toolkit 12.3.2 was used:<br>
</span><span style="font-size:11pt;font-family:"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>
+-----------------------------------------+----------------------+</span><span style="font-size:11pt"><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>
</span><span style="font-size:11pt;font-family:"Courier New"">./install_cp2k_toolchain.sh --mpi-mode=no --with-cmake=system -j 16 --enable-cuda=yes --gpu-ver=A100</span><span style="font-size:11pt"><br>
<br>
and then in the CP2K source folder:<br>
<br>
</span><span style="font-size:11pt;font-family:"Courier New"">make -j 16 ARCH=local_cuda VERSION=ssmp</span><span style="font-size:11pt"><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<u></u><u></u></span></p>
<p class="MsoNormal" style="margin-left:36pt"><span style="font-size:11pt">--
<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" target="_blank">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" target="_blank">
https://groups.google.com/d/msgid/cp2k/affc7786-0b86-4e51-8d40-22247d0d2ab9n%40googlegroups.com</a>.<u></u><u></u></span></p>
</div>
</div>
</div>
</div>


<p></p>

-- <br>
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.<br>
To unsubscribe from this topic, visit <a href="https://groups.google.com/d/topic/cp2k/wJao2K2BFMw/unsubscribe" target="_blank">https://groups.google.com/d/topic/cp2k/wJao2K2BFMw/unsubscribe</a>.<br>
To unsubscribe from this group and all its topics, send an email to <a href="mailto:cp2k+unsubscribe@googlegroups.com" target="_blank">cp2k+unsubscribe@googlegroups.com</a>.<br>
To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/cp2k/ZRAP278MB08270DCEE2C5AF6F1C0A5AE5F45A2%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer" target="_blank">https://groups.google.com/d/msgid/cp2k/ZRAP278MB08270DCEE2C5AF6F1C0A5AE5F45A2%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM</a>.<br>
</div></blockquote></div>

<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/CABdS-%3DcQ_KXs0rcPStdzVApKunwtjvat7_Vw5Sx2yg6oEM5Lxw%40mail.gmail.com?utm_medium=email&utm_source=footer">https://groups.google.com/d/msgid/cp2k/CABdS-%3DcQ_KXs0rcPStdzVApKunwtjvat7_Vw5Sx2yg6oEM5Lxw%40mail.gmail.com</a>.<br />