[gmx-users] Gromacs 2018.3 with CUDA - segmentation fault (core dumped)

[Szilárd Páll]

That suggest there is an issue related to the CUDA FFT library -- or
something else indirectly related. Can you use a newer CUDA and try to see
if with -pmefft gpu you are still getting a crash?

--
Szilárd

On Mon, Nov 12, 2018, 11:58 AM Krzysztof Kolman <krzysztof.kolman at gmail.com
wrote:

> >
> > Dear Benson and Szilard,
> >
> > Thank you for your interest. Benson, I will try to answer you questions
> > first:
> > 1) No, I have only tested 2018.3 so far. I just changed from Gromacs
> 5.1.4
> > and these are my first tests.
> > 2) Not yet but I plan to do it.
> > 3) The results look reasonable after restart
> > 4) I think so but I did not this time. My computer has 16 gb of ram and I
> > checked the ram utilization 2h before the crash. It was only using 2gb
> out
> > of 16gb.
> > 5) I use my private computer so I think it should be possible to
> recompile
> > if needed or run with debug infortmation.
> >
> > Szilard. No but they are quite close. The first crashed happened at
> > 22536600, the second one 45006200. I did earlier different simulation
> > and it also crashed (seg fault) after around 12 h.
> >
>
>    Ok. After more research I have manage to solve the problem. Mdrun with
> following flags does not trigger seg fault anymore:
>    gmx mdrun -v -deffnm md_0_1 -nb gpu -pme cpu -pmefft cpu
>
>
>
> > Best regards,
> > Krzysztof
> >
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>