Dear LAM_MPI users,
I am using LAM-MPI 7.0.4 on AMD ATHLON PC Linux cluster (8 dual CPU) with pgi and Intel compilers. After running
my code I got the next message, but code has a running status (I have seen it by "top" command), but
he does not create any output.
LAM internal GM send: gmID=3 'kappa2' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=7 'kappa5' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=4 'kappa3' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=3 'kappa2' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=8 'kappa4' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=8 'kappa4' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=4 'kappa3' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=7 'kappa5' send failed to complete (see kernel log for details): send timed out
LAM internal GM send: gmID=2 'kappa1' send failed to complete (see kernel log for details): send timed out
What doe it mean?
And another question. For my applications I use host file for lamboot:
kappa0 cpu=1
kappa1 cpu=2
kappa2 cpu=2
kappa3 cpu=2
kappa4 cpu=2
kappa5 cpu=2
kappa7 cpu=2
After lamboot, I run my code:
mpirun -np 13 ./test.x input > output &
Am I right? May be for clusters exist a better "mpirun" options? I have not any batch syestems (PBS, ..).
Thanks,
Sergey
|
