[FLASH-USERS] MPI_abort flash4

Hava Turkakin turkakin at ualberta.ca
Thu Jul 24 16:43:47 EDT 2014


Hi
I am running flash4 with a program I used for flash3 and wasa working fine.
It runs up to some point and ends with a large errro message starting with
PI_ABORT was invoked on rank 7 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------
and ending with
forrtl: error (78): process killed (SIGTERM)
Image              PC                Routine            Line        Source
libmthca-rdmav2.s  00002AAAAB0EBB69  Unknown               Unknown  Unknown
mca_btl_openib.so  00002B57E116292D  Unknown               Unknown  Unknown
libmpi.so.1        00002B57DC40E0E6  Unknown               Unknown  Unknown
libmpi.so.1        00002B57DC3409B4  Unknown               Unknown  Unknown
mca_coll_tuned.so  00002B57E264F8CE  Unknown               Unknown  Unknown
mca_coll_tuned.so  00002B57E2658528  Unknown               Unknown  Unknown
mca_coll_tuned.so  00002B57E264FC7F  Unknown               Unknown  Unknown
libmpi.so.1        00002B57DC34F97F  Unknown               Unknown  Unknown
libmpi_f77.so.1    00002B57DC0CD083  Unknown               Unknown  Unknown
flash4             00000000004397A7  Unknown               Unknown  Unknown
flash4             0000000000500700  Unknown               Unknown  Unknown
flash4             000000000044CDE9  Unknown               Unknown  Unknown
flash4             0000000000428E51  Unknown               Unknown  Unknown
flash4             000000000042FE89  Unknown               Unknown  Unknown
flash4             0000000000423D4C  Unknown               Unknown  Unknown
libc.so.6          000000367081D994  Unknown               Unknown  Unknown
flash4             0000000000423C49  Unknown               Unknown  Unknown
[cl1n126:19225] 1 more process has sent help message help-mpi-api.txt
/ mpi-abort
[cl1n126:19225] Set MCA parameter "orte_base_help_aggregate" to 0 to
see all help / error messages

                        139,1         Bot
**********************************************************************************************************
I know this about refinement. If I set lref_min=1, lref_max=6 it works
fine. But lref_min=4, lrefmax=8 dies after some point in time. I tried
to set up with usm, 8wave, increase maxblocks, all ended with the same
problem.
I want to use this refinement since resolution looks very well with
it. Any help is appreciated. Thank you
-- 
Hava Turkakin
PhD Candidate
University of Alberta
Space Physics Research Group



More information about the flash-users mailing list