HI Eugene
If you haven’t tried yet, you can set the environment variable OMP_STACKSIZE=64m.
HTH
Matthias
--
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/52bbc3d3-e857-4f21-a33f-aa42e30d106an%40googlegroups.com.
Dear Matthias,
Thank you for advice. This guide me towards localization of the
problem.
I have two testing system with the same atomic structure but different functionals, namely common PBE and hybrid PBE0. Both works in hybrid mode in CP2K v8 without changing OMP_STACKSIZE. In new version of CP2K v2024.1 it become very important to set OMP_STACKSIZE=44m. for PBE model. value below 44m will rise segfault. But for PBE0 model there is no reasonable value to finish calculations even 1024m rise segfault.
According to htop statistics I found that new 2024.1 version took more SHM (amount of shared memory used by a task) per process -- approx 1.3 times more ~ 135 M.
At the same time in case of old v8 CP2K there is no influence of OMP_STACKSIZE at all. Even very small values doesn't rise error and calculations finished normally.
Is there something related to OMP_STACKSIZE in configuration of CP2K source code. Some kind of limit or -D parameter?
I found related unresolved issue in mailing list:
https://groups.google.com/g/cp2k/c/40Ods3HYW5g
I made some test with different cutoffs as well and in new 2024.1 it always crashed in PBE0.
Best regards,
Eugene
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/TFgAsWkpnW0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZRAP278MB08278F9087D15079913994C1F42C2%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM.
Dear Eugene
The cp2k code has been increasingly functionalized with OpenMP directives since version 8 including the loops in the PW part operating with large arrays. By contrast to the GNU compiler, the Intel compiler requires now in most cases an explicit increase of the OMP_STACKSIZE. Your problem with the PBE0 run could have also other reasons. Without an input/output example showing the problem, it is difficult to provide further hints.
Best
Matthias
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/b7c23615-29ca-474d-884e-f861eef46093%40gmail.com.
Dear Matthias,
I'm really appreciate for your valuable comments. Your help is very fruitful, thank you!
Actually, where is nothing special in output
file the calculations simple freeze after the message"
HFX_MEM_INFO| Est. max. program size before HFX [MiB]: 1283
There is already a mess in different CP2K versions and supporting libraries installed, but I'm trying to control the situation...
So before sending a proper output I emerged a test using only MPI parallelization and it was always failed when hybrid functional is used. But I swear it worked before. So I grabbed dependency libraries form version 8 set and feed them to version 2024. I found libint cause the problem. According to toolchain scripts the 2.6.0 version is recommended with LMAX=5. I took the source from libint-v2.6.0-cp2k-lmax-4.tgz traball and compiled with options --enable-fortran --with-pic --enable-shared and this combination seems to work in hybrid MPI+OMP mode. Although OMP_STACKSIZE requires to be 4 times as much with respect to version 8.
Why CP2K is so sensitive to libint in Hybrid
functional calculations?
Thank you in advance,
Eugene
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZRAP278MB0827C82560F5067265202C0BF4332%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM.
Dear Eugene
As far as I know the libint tar files provided for the build of CP2K haven’t changed since quite a while (many years). I don’t know why you observe such kind of “sensitivity”.
Best
Matthias
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/7cdacb12-cb1e-4f56-9c1f-268b98e7b77c%40gmail.com.