[mpich-discuss] mpich2 : Fatal error in MPI_Comm_size: Invalid communicator

[Anthony Chan]

----- "Samir Khanal" <skhanal at bgsu.edu> wrote:

> Hi All
> 
> Does it make sense to compile by mpich-1.2.7 and execute using mpiexec
> in mpich2?

You CANNOT use mpich2's mpiexec to launch a mpich-1 compiled executable.
mpich2 is a total rewrite, it is very different from mpich-1.x.x.

> 
> My program runs well (gets compiled and gets submitted) in Mpiexec
> (OSC) 0.75 and mpich 1.2.5 GCC 4.1.1 torque 1.0.1p5 x86 gentoo
> 
> I am trying to port into 64 bit cluster (i am compiling it there) with
> GC 4.1.2, mpiexec (OSC) 0.83, mpich2(with nemesis channel) (tried
> mpich 1.2.7 and open mpi) and Torque 2.3.6

You can compile your code with mpich2's mpi wrapper compiler, mpicc/mpif90...
and use OSC's mpiexec with Torque.

> Fatal error in MPI_Comm_size: Invalid communicator, error stack:
> MPI_Comm_size(112): MPI_Comm_size(comm=0x5b, size=0x7fff4fdc906c)

The comm argument shows that you are using a mpi.h file not from MPICH2.
Most likely you are using mpi.h from mpich-1 and link with mpich2 library.
I suggest you use (full pathname) mpicc/mpif90 from mpich2 to compile and link
your application, then use OSC's mpiexec to launch your mpich2 app in Torque.  


> failed
> MPI_Comm_size(70).: Invalid communicatorrank 0 in job 30 
> comet.cs.bgsu.edu_35155   caused collective abort of all ranks
>   exit status of rank 0: killed by signal 9

A.Chan