PETSc Stokes memory issue

General Discussion
1

Hello,

I am trying to solve a Stokes problem with PETSc on an “large” mesh.
I do freefem exemple stokes-fieldsplit-3d-PETSc.edp and the exemple here.
When the mesh have 530 000 elements the use of memory is around 90% and with a mesh > 580 000 elements the use of memory is 100% and the calculation stop.

I am not familiar with PETSc and large mesh solve by freefem but I am little bit surprised by the memory usage do by PETSc.

configuration :

  • UBUNTU 20.04
  • freefem 4.11
  • 32*2 procs and 32G0 of memory

error message :

Primary job terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.

mpiexec noticed that process rank 15 with PID 0 on node tsl-ordi24 exited on signal 9 (Killed).

2

Does it work with a smaller mesh? Then either use a bigger machine, or a cheaper preconditioner to solve the problem.

3

Hello,

Yes it works with a smaller mesh (up to 530 000 tetrahedrons).
Unfortunately it is my biggest machine, 32Go of memory is not enough?
I will try with another conditioner.

Thanks you
Alexandre

4

It can be enough depending on how you precondition the problem.