[gmx-developers] Zero cell size

[Berk Hess]

Hi,

This is a bug that only shows up with certain compilers or a compiler bug.
This probably means that you will have to do some debugging.
I ran an x86_64 gcc binary through valgrind and found no memory problems.

The fatal error is generated in set_dd_ns_cell_sizes in domdec.c.
I would guess that the DD cell fractions are zero, but I don't know
why this is so.
I have committed some minor changes in domdec.c, but I don't expect
any effect of these on your problems.

You can run mdrun -debug 1 with both binaries and compare mdrun0.log.
But I don't think this will tell you much.
I expect it would also go wrong with only 2 cores.
Then you can relatively easily use a debugger to trace why the cell 
fractions
are zero.

Berk.


Roland Schulz wrote:
>
>
> On Tue, Jul 22, 2008 at 12:33 AM, David van der Spoel 
> <spoel at xray.bmc.uu.se <mailto:spoel at xray.bmc.uu.se>> wrote:
>
>     Roland Schulz wrote:
>
>         Hi,
>
>         now I get:
>         The maximum allowed number of cells is: X 0 Y 0 Z 5
>
>         Both with or without dlb. The system dimensions are the same
>         in all three dimensions.
>
>         With the old binary it still works without dlb.
>         I was using an old version of grompp but also with a current
>         grompp I get this error.
>
>         Could you send me your working tpr file? I would like to check
>         whether it is the input system our the different
>         compilation/platform.
>
>         Because this is not only with this input
>
>     Did you run the test set?
>     Anything other out of the ordinary with your installation? (e.g.
>     your own FFT routines :)).
>
>
> Sorry for making the thread confusing by replying to myself.
>
> The error: "The maximum allowed number of cells is: X 0 Y 0 Z 5" must 
> have been caused by a bad build. I couldn't reproduce it after 
> rebuilding again.
>
> The error: "The X-size (0.000000) times" seems to be caused by PGI 
> 7.1.6, because PGI 7.2 and Gcc produce correct results.
>
> My FFT isn't integrated yet ;-). (waiting on the PME refactoring by Erik)
>
> Roland
>
>  
>
>
>
>     -- 
>     David van der Spoel, Ph.D., Professor of Biology
>     Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala
>     University.
>     Box 596, 75124 Uppsala, Sweden. Phone:  +46184714205. Fax:
>     +4618511755.
>     spoel at xray.bmc.uu.se <mailto:spoel at xray.bmc.uu.se>  
>      spoel at gromacs.org <mailto:spoel at gromacs.org>  
>     http://folding.bmc.uu.se
>
>     _______________________________________________
>     gmx-developers mailing list
>     gmx-developers at gromacs.org <mailto:gmx-developers at gromacs.org>
>     http://www.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
>     <mailto:gmx-developers-request at gromacs.org>.
>
>
> This email was Anti Virus checked by Astaro Security Gateway. http://www.astaro.com
> ------------------------------------------------------------------------
>
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.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.