Crashes

From Henkelman Group

Jump to: navigation, search

A page to document crash reports


[edit] Ranger

Here is output during job initialization TACC: Setting up parallel environment for MVAPICH-2 MPD. running mpdallexit on i130-107.ranger.tacc.utexas.edu .... Here is the actual error written to output file.

Fatal error in MPI_Isend: Internal MPI error!, error stack: MPI_Isend(163): MPI_Isend(buf=0xd138450, count=796, MPI_DOUBLE_PRECISION, dest=44, tag=201, comm=0xc4000004, request=0x7fff045ade18) failed (unknown)(): Internal MPI error! rank 144 in job 1 i130-107.ranger.tacc.utexas.edu_45384 caused collective abort of all ranks exit status of rank 144: killed by signal 9 TACC: MPI job exited with code: 137 TACC: Shutting down parallel environment. TACC: Shutdown complete. Exiting. TACC: Cleaning up after job: 133733 TACC: Done.

Response from TACC Consulting:

I see that you are using mvapich2. Do you really need the functionality of mvapich2. If not, please retry with the mvapich/1.0 or mvapich-devel/1.0. These stacks work much better. see Ranger_NEB_error for more information and a second attempt's output Markm 15:30, 25 June 2008 (CDT)

Retrieved from "http://theory.cm.utexas.edu/wiki/index.php/Crashes"
Personal tools