Arham,
It should be balanced: the default mapping is to allocate NUMA packages
round robin.
you can
mpirun --report-bindings -n 28 true
to have Open MPI report the bindings
or
mpirun --tag-output -n 28 grep Cpus_allowed_list /proc/self/status
to have each task report which physical cpu it is
Hi
If I run a code with
mpirun -n 28 ./code
Is it guaranteed that Open MPI and/or OS give equal number of processes to
each socket? Or I have to use some mpirun options?
Running the code with the command given above, one socket gets much hotter
than the other (60°C vs 80°C). I'm sure that the
