Fix #914

[wilfonba]

User description

Description

This fixes the bug with boundary condition patches in post process.

Fixes #914

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

Scope

• This PR comprises a set of related changes with a common goal

If you cannot check the above box, please split your PR into multiple PRs that each have a common goal.

How Has This Been Tested?

Run examples/2D_jet/case.py with 4 ranks. Output and video is attached.
out.txt

test.mp4

---

PR Type

Bug fix

---

Description

• Fix boundary condition patches in post process

• Add missing bc_io variable to MPI broadcast list

---

Changes diagram

flowchart LR
  A["MPI broadcast variables"] --> B["Add bc_io variable"] --> C["Fix 2D_jet post process"]

Loading

---

Changes walkthrough 📝

	Relevant files
Bug fix	m_mpi_proxy.fpp Add missing bc_io to MPI broadcast                                              src/post_process/m_mpi_proxy.fpp Add bc_io variable to the MPI broadcast variable list Fix missing variable causing post process to hang +1/-1

---

Need help?

Type /help how to ... in the comments thread for any questions about Qodo Merge usage.

Check out the documentation for more information.

[Copilot]

Pull Request Overview

This PR adds support for the new bc_io boundary-condition flag in the MPI proxy broadcast routine.

• Included bc_io in the list of flags to be broadcast
• Ensures that post-process modules receive the updated boundary-condition setting

Comments suppressed due to low confidence (1)

src/post_process/m_mpi_proxy.fpp:110

• Consider adding a focused test (unit or integration) to verify that the bc_io flag is correctly broadcast and applied across all ranks in the MPI proxy.

            & 'output_partial_domain', 'relativity', 'cont_damage', 'bc_io']

[qodo-code-review]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

🎫 Ticket compliance analysis ✅ 914 - PR Code Verified Compliant requirements: • Fix the 2D_jet example that gets stuck during post process • Resolve the hanging issue that prevents the example from completing Requires further human verification: • Verify that the 2D_jet example now runs successfully without hanging • Confirm that other examples still work correctly after this change

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Missing Tests The fix adds a missing variable to MPI broadcast but lacks automated tests to prevent regression. Consider adding a test that verifies all required variables are properly broadcast in MPI operations. & 'output_partial_domain', 'relativity', 'cont_damage', 'bc_io'] call MPI_BCAST(${VAR}$, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)

[qodo-code-review]

PR Code Suggestions ✨

No code suggestions found for the PR.

[sbryngelson]

This closes #914 but opens (please open @wilfonba) an issue that we are seemingly not testing correctly. Can you figure out why this example case was passing when it should have been failing?

[wilfonba]

This closes #914 but opens (please open @wilfonba) an issue that we are seemingly not testing correctly. Can you figure out why this example case was passing when it should have been failing?

The test suite doesn't test multiple rank cases very thoroughly. It only has 3 MPI cases in the no RDMA case.

[sbryngelson]

This closes #914 but opens (please open @wilfonba) an issue that we are seemingly not testing correctly. Can you figure out why this example case was passing when it should have been failing?

The test suite doesn't test multiple rank cases very thoroughly. It only has 3 MPI cases in the no RDMA case.

Was that the actual problem? This problem occurred for a single rank CPU case on my laptop...

[wilfonba]

This closes #914 but opens (please open @wilfonba) an issue that we are seemingly not testing correctly. Can you figure out why this example case was passing when it should have been failing?

The test suite doesn't test multiple rank cases very thoroughly. It only has 3 MPI cases in the no RDMA case.

Was that the actual problem? This problem occurred for a single rank CPU case on my laptop...

I just ran it with one rank on the master branch on my laptop, and it ran fine. It only got stuck in post process when I added a second rank. I'm not sure what you did differently on your end. The issue is that when bc_io isn't broadcast, the non-root processes don't attempt to open an MPI_IO file, and a gridlock occurs.

[sbryngelson]

I'm not sure what I did either, I guess we will just proceed.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 44.03%. Comparing base (f1fea7a) to head (1c3cd99).
Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #916   +/-   ##
=======================================
  Coverage   44.03%   44.03%           
=======================================
  Files          68       68           
  Lines       18395    18395           
  Branches     2227     2227           
=======================================
  Hits         8101     8101           
  Misses       8991     8991           
  Partials     1303     1303           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.