On Apr 15, 2005, at 2:59 PM, Jeff Polasek wrote:
> I have been fighting for several days with trying to get LAMMPI to
> work with GM 2.0.19 (myrinet) on our G5 (OS X 10.3.7) cluster. I have
> been through the archives, found many possible solutions, but still
> don't have it working.
> [snip]
> Running laminfo gives the following
>
> node065:/usr/local jjp8508$ laminfo
> LAM/MPI: 7.2b1svn10122
> [snip]
> SSI rpi: gm (API v1.1, Module v1.2)
> [snip]
> node065:~/gm jjp8508$ mpirun C -ssi rpi gm -ssi rpi_verbose level:1000
> ./cpi
> [snip]
> n0<16758> ssi:rpi:open: looking for rpi module named gm
> -----------------------------------------------------------------------
> ----The rpi module named "gm" could not be found.
Wow -- that's a new one for me! :-(
It's quite odd that laminfo shows the module but then mpirun'ing
doesn't find it. I'm really not quite sure how that could happen.
I realize that you said that you double checked your path to ensure
that you're getting the right mpirun, but:
- can you triple check it? In my experience, seeing these kinds of
"this should never happen" kinds of errors is when you're not running
what you think you're running
- also check the output of "mpicc -showme" to see that it's showing the
right directories for header files and libraries
- check to ensure that the LAMHOME environment variable is *not* set --
this is something that can throw off mpicc (and others), even if you're
using the "right" mpicc, etc.
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
