Page 2 of 2
Re: Coupling Neptune/syrthes
Posted: Mon Jul 18, 2016 8:59 am
by fchrit
Hello,
Are there any news please?
Fatima
Re: Coupling Neptune/syrthes
Posted: Fri Jul 22, 2016 8:49 am
by fchrit
Hello again,
I got some news, I now get this new error message but I still don't know how to get through it:
[chausson:10825] *** An error occurred in MPI_Allreduce
[chausson:10825] *** reported by process [47672487641089,47669842018304]
[chausson:10825] *** on communicator MPI_COMM_WORLD
[chausson:10825] *** MPI_ERR_COMM: invalid communicator
[chausson:10825] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[chausson:10825] *** and potentially your MPI job)
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI COMMUNICATOR 3 SPLIT FROM 0
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.
--------------------------------------------------------------------------
[chausson:10823] 1 more process has sent help message help-mpi-api.txt / mpi-abort
[chausson:10823] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
solver script exited with status 5.
Can anyone help please?
Fatima
Re: Coupling Neptune/syrthes
Posted: Fri Jul 22, 2016 9:51 am
by Yvan Fournier
Hello,
Did you check the listing (fluid) and syrthes.log (solid) files ?
Regards,
Yvan
Re: Coupling Neptune/syrthes
Posted: Fri Jul 22, 2016 10:28 am
by fchrit
Yes, the second fluid listing gives this error :
SYRTHES coupling 0: initializing MPI communication ... SIGTERM signal (termination) received.
--> computation interrupted by environment.
Syrthes can't read the second coupling I think but I can't figure out why?
Fatima
Re: Coupling Neptune/syrthes
Posted: Fri Jul 22, 2016 11:52 am
by fchrit
Hey again,
I changed the names of the fluid cases and now syrthes can read the 2 surfaces. However, it stops after the 1st iteration and gives this error in syrthes.log file :
[chausson:27801] *** Process received signal ***
[chausson:27801] Signal: Segmentation fault (11)
[chausson:27801] Signal code: Address not mapped (1)
[chausson:27801] Failing at address: (nil)
[chausson:27801] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x10330)[0x2b29e8ba0330]
[chausson:27801] [ 1] /lib/x86_64-linux-gnu/libc.so.6(fclose+0x4)[0x2b29e8e1b8d4]
[chausson:27801] [ 2] ./syrthes[0x40767f]
[chausson:27801] [ 3] ./syrthes[0x402cbf]
[chausson:27801] [ 4] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x2b29e8dcff45]
[chausson:27801] [ 5] ./syrthes[0x402ee5]
[chausson:27801] *** End of error message ***
and in the terminal I get this:
mpiexec noticed that process rank 0 with PID 29427 on node chausson exited on signal 11 (Segmentation fault).
Can anyone help please?
Fatima
Re: Coupling Neptune/syrthes
Posted: Fri Jul 22, 2016 12:15 pm
by Yvan Fournier
Hello,
There are recent threads on this forum in which I explain how to compile Syrthes in debug mode, so you can try that. It may provide a bit more info on where the issue is coming from, so if it is a known issue (for which patches should also be available on somewhat older discussion threads here), you may solve it. If it is an "unknown" Syrthes bug, someone would need to debug that...
Regards,
Yvan