Hello Xi
I will have to get back on the SiCortex to take a look at this- I
will get back to you early this week. I have run this on a SiCortex but
it has been a while...
Justin
On Fri, 11 Feb 2011, Xi Duan wrote:
> Hi all,
> I followed the the quick start on the home page
> (http://c-mpi.sourceforge.net/), and it was all good
> when I was doing the setup and configure steps with mpich2-1.2.1p1, and my
> machine is Sicortex072. But I met problems while the make step.
>
> At the beginning, it turned out the error of
> "//home/xduan/mpich2-1.2.1p1-install/bin/mpdroot: open failed for root's mpd
> conf filempdlistjobs (__init__ 1208): forked process failed; status=255/".
> Then I followed the discuss threads at
> http://lists.mcs.anl.gov/pipermail/mpich-discuss/2010-March/006782.html, and
> set /mpd.conf /file at both//etc// and//usr/etc//, moreover, I didn't include
> the /"MPD_USE_ROOT_MPD=1" /thing due to I just use a regular user started
> ring. But it returned the error at /"test/mpirpc/test-ping.out.failed"/,
> which says /"srun: error: Unable to allocate resources: No partition
> specified or system default partition"/.
>
> So I realized I need to run multiple instance instead of just single process.
> Then I made the ring and used "srun" to start multiple processes on different
> machines. However, I still get the error message at
> /"test/mpirpc/test-ping.out.failed"/, the following paragraph is the error
> code:
>
> /hello
> hello
> Using single debug file
> [0] debug_rank: 0
> Fatal error in MPI_Isend: Invalid rank, error stack:
> MPI_Isend(146): MPI_Isend(buf=0x1200dc3c8, count=1, MPI_INT, dest=1, tag=0,
> MPI_COMM_WORLD, request=0x120115c30) failed
> MPI_Isend(96).: Invalid rank has value 1 but must be nonnegative and less
> than 1
> Using single debug file
> [0] debug_rank: 0
> Fatal error in MPI_Isend: Invalid rank, error stack:
> MPI_Isend(146): MPI_Isend(buf=0x1200dc3c8, count=1, MPI_INT, dest=1, tag=0,
> MPI_COMM_WORLD, request=0x120115c30) failed
> MPI_Isend(96).: Invalid rank has value 1 but must be nonnegative and less
> than 1
> srun: error: task 0-1: Exited with exit code 1
> un: error: task 1: Exited with exit code 1/
>
> Now I am stopped by this problem, could anyone provide any idea? BTW, I also
> tried a root-started ring with /"MPD_USE_ROOT_MPD=1" /in two locations.
> Nevertheless, they have the same problem.
>
> Looking forward to your responses.
>
> Thanks,
> Xi
>
--
Justin M Wozniak
|
