From: Aron Broom (broomsday_at_gmail.com)
Date: Fri Apr 20 2012 - 09:25:40 CDT
Peter,
I've seen similar problems with NAMD 2.8, that is, constraint failures on
the GPU that don't show up on the CPU. I don't think it is a problem with
NAMD, however. It think there are likely two possible culprits:
1) You really do have some constraint problems (system is at high energy,
your temperature and timestep are not appropriate for using contraints, so
forth and so on).
2) Your GPU is slightly faulty. I mainly noticed these problems when using
an nVidia GTX card that I later found to have memory problems. If you
google OpenMM memtest, you can find a little command-line tool for testing
your GPU memory, might be a worthwhile sanity check.
~Aron
On Fri, Apr 20, 2012 at 1:45 AM, Norman Geist <
norman.geist_at_uni-greifswald.de> wrote:
> Hi Peter,
>
> Does the same inputs work with another namd version?
>
> Best wishes
>
> Norman Geist.
>
> > -----Ursprüngliche Nachricht-----
> > Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im
> > Auftrag von Peter Jones
> > Gesendet: Donnerstag, 19. April 2012 11:31
> > An: namd-l_at_ks.uiuc.edu
> > Betreff: namd-l: problem with gpu namd 2.9b3
> >
> > Hello,
> >
> > I'm trying to run a simulation using Namd 2.9b3 on a cluster with gpu
> > acceleration; its a continuation of a job that was running on another
> > machine. Namd starts ok, prints out the energies for step 0, and then
> > prints-
> >
> > LDB: ============= START OF LOAD BALANCING ============== 11.8193
> > LDB: ============== END OF LOAD BALANCING =============== 11.8194
> > LDB: =============== DONE WITH MIGRATION ================ 11.8197
> > 0:961(5) atom 32894 found 2 exclusions but expected 3
> > 0:1042(50) atom 40634 found 2 exclusions but expected 3
> >
> > followed by a long list of messages similar to the last two lines. This
> > is followed by a list of errors regarding RATTLE constraint failures
> > and then termination. I'm new to gpu-accelerated NAMD and I do not have
> > any idea what is wrong here, any help would be greatly appreciated,
> >
> > Thankyou,
> >
> > Kind Regards,
> > Peter Jones
> >
>
>
>
>
-- Aron Broom M.Sc PhD Student Department of Chemistry University of Waterloo
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:27 CST