[gmx-developers] GB Dom Decomp problem

[David van der Spoel]

On 2010-08-06 22.43, hess at sbc.su.se wrote:
> Hi,
>
> This should not happen.
> Did you get some information on the grid size/shift somewhere?
>
Not other than that reported below. Each of the 8 cores gives this 
message for its specific cell. The molecule is fixed in the center of 
the box and stays there (as checked with g_traj).

Continuing from a checkpoint file solves the problem.

Anything else I can do to debug it?
> Berk
>
>> Hi,
>>
>> I'm running a long GV simulation on 8 cores, and after tens of
>> nanoseconds it crashes with this error:
>>
>> Program mdrun, VERSION 4.5-beta1-20100730-f9ac35b
>> Source code file: /scratch/spoel/GG/gromacs/src/mdlib/domdec.c, line: 2629
>>
>> Fatal error:
>> Step 83099555: The domain decomposition grid has shifted too much in the
>> Y-direction around cell 1 0 0
>>
>> Original output is:
>>
>> Making 3D domain decomposition 2 x 2 x 2
>> starting mdrun 'Generated by trjconv : MPRA t= 2000.00000'
>> 100000000 steps, 200000.0 ps (continuing from step 82508605, 165017.2 ps).
>> step 82508605, will finish Tue Aug 10 16:01:41 2010
>>
>> Simulations run with angular comm removed, so the protein should not
>> move away from the center of the box.
>>
>> Can this be prevented?
>>
>>
>> --
>> David van der Spoel, Ph.D., Professor of Biology
>> Dept. of Cell&  Molec. Biol., Uppsala University.
>> Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
>> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
>> --
>> gmx-developers mailing list
>> gmx-developers at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-developers
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-developers-request at gromacs.org.
>>
>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se