[mpich-discuss] how to make mpich crash hard so I can gdb bt it?

Jeff Hammond jeff.science at gmail.com
Fri Oct 11 16:44:52 CDT 2013


I apologize if this is a stupid question, but I would like MPICH to
crash and burn rather than terminate gently when an error occurs so
that I can gdb bt it to know where the error is in my program, since I
am absolutely certain that this error is in the application and not
MPICH.

Ideally, I would be able to select the failure mode at runtime, e.g.
MPICH_FAILURE_MODE={hard,soft}, so that I can get either the nice
MPICH trace when I think MPICH is the problem and a brutal
light-the-machine-room-on-fire-and-abort-the-universe error when I
think it's my fault.

Thanks,

Jeff

<running in gdb>
Assertion failed in file ../src/mpi/coll/helper_fns.c at line 81: FALSE
memcpy argument memory ranges overlap, dst_=0x101401000
src_=0x101401000 len_=400

internal ABORT - process 0
Fatal error in MPI_Win_free: Other MPI error, error stack:
MPI_Win_free(134)..........: MPI_Win_free(win=0x10000b394) failed
MPIDI_CH3_SHM_Win_free(103): pthread mutex routine failed (Resource busy)

Program exited with code 01.
(gdb) bt

$ mpichversion
MPICH Version:    	3.1b2
MPICH Release date:	unreleased development copy
MPICH Device:    	ch3:nemesis
MPICH configure: 	CC=clang CXX=clang++ FC=/bin/false --disable-fc
--disable-f77 --enable-threads --enable-g --with-pm=hydra
--prefix=/opt/mpich/dev/clang
MPICH CC: 	clang    -g -O2
MPICH CXX: 	clang++   -g -O2
MPICH F77: 	no   -g
MPICH FC: 	no   -g

-- 
Jeff Hammond
jeff.science at gmail.com



More information about the discuss mailing list