[gmx-users] Signal:11 info.si_errno:0(Success) si_code:1(SEGV_MAPERR)

[David van der Spoel]

Mike Hanby wrote:
> Howdy, a user of a cluster that I manage sent me an output file of their
> Gromacs 3.3.1 job. It looks like the job runs for approx 11 hours and
> then abruptly terminates with a SEGV_MAPERR. What could cause this
> (gromacs, or a system config / resource?).
> 
> The output of the log file is below, thanks for any suggestions. Mike
> 
> ########################################################################
> #############
>  STARTED AT: Wed Jan 17 11:40:03 CST 2007
> 
> NSLOTS: 16
>   :-)
> /share/apps/gromacs/intel/gromacs-3.3.1-s64-openmpi/bin/mdrun_ompi  (-:
> 
> Option     Filename  Type         Description
> ------------------------------------------------------------
>   -s production-Npt-300K_16CPU.tpr  Input        Generic run input: tpr
> tpb
>                                    tpa xml
>   -o production-Npt-300K_16CPU.trr  Output       Full precision
> trajectory:
>                                    trr trj
>   -x production-Npt-300K_16CPU.xtc  Output, Opt! Compressed trajectory
>                                    (portable xdr format)
>   -c production-Npt-300K_16CPU.gro  Output       Generic structure: gro
> g96
>                                    pdb xml
>   -e       ener.edr  Output       Generic energy: edr ene
> 
> Getting Loaded...
> Reading file production-Npt-300K_16CPU.tpr, VERSION 3.3.1 (single
> precision)
> Loaded with Money
> 
> Back Off! I just backed up ener.edr to ./#ener.edr.1#
> starting mdrun 'Protein in water'
> 2500000 steps,   2500.0 ps.
> 
> ...... steps ......
> 
> step 650170, will finish at Fri Jan 19 06:52:24 2007
> 
> step 650180, will finish at Fri Jan 19 06:52:21 2007
> Signal:11 info.si_errno:0(Success) si_code:1(SEGV_MAPERR)
> Failing at addr:0x7c3fefc
> [0] func:/share/apps/openmpi/intel/openmpi-1.1.2-64/lib/libopal.so.0
> [0x2a966102a5]
> [1] func:/lib64/tls/libpthread.so.0 [0x3793b0c430]
> [2]
> func:/share/apps/gromacs/intel/gromacs-3.3.1-s64-openmpi/bin/mdrun_ompi(
> spread_q_bsplines+0x1c0) [0x45c80c]

We've seen quite some problems with MPI libraries previously. I don't 
know what the current status of openmpi is. As the cluster manager you 
may have run some mpi tests, if so did they work flawless?

The gromacs function (2 in your call tree) could also generate an 
overflow due to physics like large forces. Hence it is not possible to 
determine from this output what causes the problem.

-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++