[CIG-SEISMO] SPECFEM3D Cartesian MPI failure due to gatherv_all_cr() in fault_solver_dynamic.f90

[Junwei Huang]

Hi all,
I compiled the development version using gcc, gfortran and openmpi and
tried the TPV5 example. I succeeded in using a single processor. With more
than 2 processors, I got MPI error messages as follows:

Operating system error: Cannot allocate memory
Memory allocation failed
--------------------------------------------------------------------------
mpirun has exited due to process rank 1 with PID 21189 on
node ubuntu exiting without calling "finalize". This may
have caused other processes in the application to be
terminated by signals sent by mpirun (as reported here).
--------------------------------------------------------------------------

It seems to me that it is the function call of gatherv_all_cr()  in either
init_dataXZ(bc%dataXZ,bc) or gather_dataXZ(bc). One processor becomes dead
after the call and the above error message will be thrown once
synchronize_all() or MPI_BARRIER(my_local_mpi_comm_world,ier) is called.

I would be really grateful if some experts here can help me figure out the
issue. Could it be the mpi library? I am using openmpi1.3. Or the example
itself needs update? Thanks very much.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20150310/8b3a616c/attachment.html>