Skip to content
Snippets Groups Projects

Fix the exit iteration padding.

Merged James Willis requested to merge fix_pair_vec_exit_iteration into master
  • Removes padding for the exit_iteration when looping over cj.
  • Calculates the padding correctly for the exit_iteration when looping over ci

Merge request reports

Merged by avatar (May 30, 2025 6:17am 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
  • Can you explain briefly why this is necessary?

  • I was incorrectly using the exit_iteration to check whether the number of particles to loop over was a multiple of VEC_SIZE. I forgot that exit_iteration was an index.

  • Ok. Is there a reason why it did not show up as a bug in the test suite? Or when pre-filling the caches with random garbage?

  • Yes, because I pad a whole vector worth of fake particles in the cache at the end of it.

  • mentioned in commit 799d3b29

  • Running ./testPeriodicBC -n 6 -r 1 -d 0 -f standard -v 0
    ./testPeriodicBC.sh: line 3: 51489 Segmentation fault      (core dumped) ./testPeriodicBC -n 6 -r 1 -d 0 -f standard -v $v
    Error Missing test output file
    FAIL: testPeriodicBC.sh
    
    Running ./testPeriodicBC -n 6 -r 1 -d 0.1 -f perturbed -v 0
    ./testPeriodicBCPerturbed.sh: line 3: 51587 Segmentation fault      (core dumped) ./testPeriodicBC -n 6 -r 1 -d 0.1 -f perturbed -v $v
    Error Missing test output file
    FAIL: testPeriodicBCPerturbed.sh
  • I think I've found the problem. I thought I was fixing a problem that didn't exist in the ci loop. exit_iteration_align needs to be a multiple of VEC_SIZE not the number of particles looped over. I will just reverse that change and it should be fine...

Please register or sign in to reply
Loading