MetOffice / MetOffice/lfric_apps

The combination of threading and MPI regions that cross cubed-sphere panel edges is not bit-reproducible

Open
#294 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug
Dominant language
Fortran
Stars
31
Forks
118
Avg merge
2d 4h
Merged PRs (30d)
8

Description

Are there any linked Issues or Pull Requests?
What happened?

It was found that after #138 was merged, that the gungho_model_baroclinic-C48_MG-3panel_azspice_gnu_fast-debug-64bit would sporadically produce different KGOs.

This will be avoided for now on main by setting the test to use a single thread (#293) but this suggests that tests which use both threading and have MPI regions that cross cubed-sphere panel edges are not bit-reproducible. My best hypothesis for now is that it relates to https://github.com/MetOffice/lfric_core/issues/252

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the gungho_model_baroclinic-C48_MG-3panel_azspice_gnu_fast-debug-64bit test configuration and review linked issues #252 and #293. Reproduce the run with threading enabled and MPI regions crossing cubed-sphere panel edges, then identify and resolve the source of differing KGOs so repeated runs are bit-reproducible.

Written by the indexing model from the issue text.

Assessment

Tech stack
fortran
Domain
distributed-systems, testing-qa
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.