Skip to content
Snippets Groups Projects

Ignore inhibited particles in timestep

Merged Matthieu Schaller requested to merge ignore_inhibited_particles_in_timestep into master

This should supersede !843 (closed).

The old test that crashed was mpirun -np 8 ../../swift_mpi --pin --cooling --star-formation --feedback --stars --cosmology --hydro --self-gravity -v 1 --threads=14 eagle_25.yml with the EAGLE subgrid model activated.

This now runs smoothly with the debugging checks activated.

Merge request reports

Merged by avatar (May 31, 2025 6:41am UTC)

Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Matthieu Schaller mentioned in merge request !843 (closed)

    mentioned in merge request !843 (closed)

  • Hmm, so you think that the old code gave the wrong value for the inhibited counts?

  • The old code was collecting the inhibited counts in the time-step function. That's not a valid approach any more with black holes or actually in jacob's runs as we sometimes remove particles in cells that are not active. As long as we only had star formation we were fine as this was happening in active cells. It was wrong for Jacob though.

    The previous MR may have been broken as I was only half-way through that transition.

    We don't want to collect the inhibited counts here any more and indeed the cells do no carry such counts any more.

    The other aspect is that we don't want to compute a time-step at all for the inhibited particles, which is the main change here.

    As far as I can tell this now runs fine on the usual eagle boxes over mpi, which is my usual main worry.

  • Thanks, so we'll accept and see how that goes this time...

  • Peter W. Draper mentioned in commit dcf71f9e

    mentioned in commit dcf71f9e

Please register or sign in to reply
Loading