Dear all
I am using a in house parallel fortran program which uses lam-mpi libraries.
The OS and compilers we are using are Fedora core4 and ifort.
This prog. successfully runs on single and duel processor machines with out
any problem
I have a prob in running the same code in duel core machines.
When i give 1 cpu per/node then the prog works fine. If i give 2cpu/node
then i will get only the one cpu output. I didn't get the results from the
second cpu.
One more thing, the job is not killed. It is running but i didn't get any
outputs.
The same prog runs well in single core machines with 5cpus/node. The job
splitted into five processes on the same node and works well.
Thanks in advance for your help and Suggestions.
Regards
Sabai.
_________________________________________________________________
Sexy, sultry, sensuous. - see why Bipasha Basu is all that and more. Try MSN
Search http://server1.msn.co.in/Profile/bipashabasu.asp
|
