Hello,

The situation is that I have an MPI-parallel fortran program. I run it and it's distributed on N processors, and each of these processes must call an external program.

This external program is also an MPI program, however I want to run it in serial, on the processor that is calling it, as if it were part of the fortran program. The fortran program waits until the external program has completed, and then continues.

The problem is that this external program seems to run on any processor, and not necessarily the (now idle) processor that called it.

Can anyone tell me how I can call the program and ensure it runs only on this processor?

Thanks!

---

Extra information that might be helpful:

If I simply run the external program from the command line (ie, type "/path/myprogram.ex <enter>"), it runs fine. If I run it within the fortran program by calling it via

CALL SYSTEM("/path/myprogram.ex")

it doesn't run at all (doesn't even start) and everything crashes. I don't know why this is.

If I call it using mpiexec:

CALL SYSTEM("mpiexec -n 1 /path/myprogram.ex")

then it does work, but I get the problem that it can go on any node.

Sounds like "processor affinity":

http://www.open-mpi.org/faq/?categor...uild-paffinity
http://www.open-mpi.org/faq/?categor...paffinity-defs
http://www.open-mpi.org/faq/?categor...sing-paffinity

1 members found this post helpful.

Hi paulsm4,

Yes, thank you!
It led me, in an indirect way, to the answer.

Using SGE and MVAPICH2, the open-mpi-specified paffinity option that you linked to doesn't seem to exist, however a slight variation on this does, using the "-env MV2_CPU_MAPPING 0:1....." option in mpiexec.

So when calling the external program with mpiexec, I map the called process to the current core rank, and it seems to stay distributed and separated as I want.

Thanks again for the hint!