[CP2K-user] [CP2K:16459] Wavelet Solver Issues

Augustin Bussy augustin.bussy at chem.uzh.ch
Fri Jan 14 12:22:10 UTC 2022


Dear Moe,

your inputs are correct, although not optimal (more on this later). I 
did some testing, and it seems that  you ran into a bug of the wavelet 
solver when large boxes and high number of OMP threads are involved. We 
will look into that.

A box of 50A x 50A x 50A is very, very large. For optimal performance, 
one should pick a box size that is just big enough, such that the 
electronic density vanishes at its edges. In your case, 18A x 18A x 18A 
is sufficient. This choice should speed things up significantly, and 
keep you away from the bug I mentioned earlier.

As a side note, CP2K is better optimized for MPI parallelization rather 
than OMP. From your output file, I can see that you run with 1 MPI rank 
and 36 OMP threads. I would suggest something like 9 MPI ranks and 4 OMP 
threads instead (I get a speedup of 10 that way).

Best,

Augustin

On 1/13/22 21:49, Moe El wrote:
> Hi Everyone,
> I am trying to run a static energy and geometry optimization of an rgd 
> chain; however, I always keep getting an error related to the solver. 
> I followed the tutorial online and tried to use the same methodology 
> for my molecule. I am very new to CP2K and I dont know what I am doing 
> wrong. I would like any feedback for improvement. Thank you all for 
> your time and effort.
> -- 
> 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/f9c3b3dd-f39b-496b-8c1d-aa5acd2d313an%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/f9c3b3dd-f39b-496b-8c1d-aa5acd2d313an%40googlegroups.com?utm_medium=email&utm_source=footer>.

-- 
Augustin Bussy
Postdoc Researcher
--------------------------

University of Zurich
Augustin Bussy
Department of Chemistry
Research Group Hutter
Winterthurerstrasse 190
CH-8057 Zurich

www.chem.uzh.ch
augustin.bussy at chem.uzh.ch
--------------------------

-- 
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/89cb1af4-1852-28bb-7047-e9313e7ffbf5%40chem.uzh.ch.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20220114/655e9788/attachment.htm>


More information about the CP2K-user mailing list