[mpich-discuss] I have a question

[michael bane]

On Tue, 2010-09-21 at 16:07 -0600, sc Deng wrote:
> Dear all,
> 
>   I meet a question in using MPICH2, the error is as follows,
> 
> 
> TimeStep =    3500  Time = 0.1050E+04
>   Total E =  85.3849     Pot E =  81.9326     Kin E =  3.45222    
>   Temp = 0.702334E-04 Pressure =-0.789195E-04 Volume =  191110.    
>   Box =  57.7138      57.4711      57.6176    
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image              PC                Routine            Line
> Source             
> DEM_Linux.out      000000000046FAE6  Unknown               Unknown
> Unknown
> 
> Stack trace terminated abnormally.
> rank 1 in job 5  fn-148-148_54121   caused collective abort of all
> ranks
> exit status of rank 1: killed by signal 9 
> 
> 
> It is a 3D Distinct element code, the error appears after computing
> 3500 time steps, but it is OK for 2D simulation or just 1 processor
> for 3D. Please let me know if anyone has the corresponding experience.
> Thank you very much.

compile with -g (and -traceback if ifort, else similar) then run again
within a debugger... you'll find it's something like out-of-bounds array
accessing

it's not an mpich2 bug

M
-- 
Dr. Michael K. Bane
Snr Research Apps & Collabs Consultant
RESEARCH COMPUTING SERVICES, ITS
Univ. of Manchester M13 9PL

http://www.rcs.manchester.ac.uk
http://www.twitter.com/mkbane_mcr