As a first time user of CosmoMc I ran it using
mpirun -np 32 ./cosmomc test.ini
where test.ini is provided with the distribution.
and as I write it is running for 48 wall-hours (so about 48 x 32 CPU hours)
When I logged in to the compute node and did an strace on one of the jobs I see
"write(23, "\f\0\0\0\10\0H\0010\263y\7\374\177\0\0", 16) = -1 EAGAIN (Resource temporarily unavailable)"
Any idea what's going on? Is this an MPI problem or I/O error? If the latter why hasn't the program crashed?
That said, the chains subdirectory is regularly getting updated with test_*.txt and test_*.data.
So not sure if its fine.
Anyhow would appreciate any help if I should wait or given the strace output, something is wrong and if so, what.
cosmoMC test jn running for a long time
1 post • Page 1 of 1