On Jul 31, 2013, at 6:45 PM, meng wrote:
> I install openmpi at /usr/local/openmpi-1.6.5 on both computers. and
> the two computers now can access each other without passwd required. I set
> the path and ld_library_path in .bashrc and source it.
> I still get the problems as before.
Dear Dani and Reuti,
>> either install openmpi on each node, and setup
>> /etc/profile.d/openmpi.{c,}sh and /etc/ld.so.conf.d/openmpi.conf files on
>> both (preferred) or install to a common file system (e.g. nfs mount) and
>> still use profile and ldconfig to setup environment. >
I cho
Sorry for this.
This was an try and ERROR ERROR Problem.
It was a mismatch of OFED versions and kernel updates.
Now I installed a fresh centOS 6.4 (with default kernel NO KENELUPDATE).
Then installed the official MELLANOX OFED Driver and compiled openMPI
(without options). And now it works fine.
Hi,
What OFED vendor and version do you use?
Regards
M
On Tue, Jul 30, 2013 at 8:42 PM, Paul Kapinos wrote:
> Dear Open MPI experts,
>
> An user at our cluster has a problem running a kinda of big job:
> (- the job using 3024 processes (12 per node, 252 nodes) runs fine)
> - the job using 4032 p
