I'm using oneapi 2021.1.1 on centos7.6.
I wanna profiling mpi application with aps but it works when I run wrf but fails when I run lammps.
So I tries to run the application locally.
When I run wrf independently without aps `./wrf`, there is normal output from wrf and it runs correctly. With aps `aps ./wrf`, there is normal output as well.
When I run lammps indenpendently without aps `./lmp -sf opt -in in.lj` , I can get normal output of lammps and it runs correctly.
When I run lammps indenpendently with aps `aps ./lmp -sf opt -in in.lj` , the output stucks at the start. I can only get:
--------------------------------------------------------------------------
WARNING: There is at least non-excluded one OpenFabrics device found,
but there are no active ports detected (or Open MPI was unable to use
them). This is most certainly not what you wanted. Check your
cables, subnet manager configuration, etc. The openib BTL will be
ignored for this job.
Local host: 6248r-node125
--------------------------------------------------------------------------
LAMMPS (5 Jun 2019)
There are vtune and aps processes:
root 40684 31718 0 15:36 pts/49 00:00:00 aps ./lmp -sf opt -in in.lj
root 40688 40684 0 15:36 pts/49 00:00:00 /opt/compiler/intel/oneapi/vtune/2021.1.1/bin64/vtune -q -collect mps -mrte-mode=native -data-limit=0 -finalization-mode=none -r /tmp/aps_temp_1804289392masterResult -- ./lmp
root 40715 40710 0 15:36 pts/49 00:00:00 ./lmp
But I cannot see any log from lammps. Normally it will run about 3~5s. But with aps It stucks at start least 1min and I didn't wait for its ending and killed it.
