NWChemEx / NWChemEx/TensorWrapper
Running unit tests in parallel
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 0
- Forks
- 0
- Avg merge
- 5h 33m
- Merged PRs (30d)
- 6
Description
I remember we had a discussion about it before but I couldn't find the issue. The problem is when you run unit tests in parallel with mpirun most of them fail with Segmentation violation signal. For PluginPlay and ParallelZone we don't have this problem, so the issue is TW specific.
I encounter this while trying to understand the problem with passing an mpi communicator to madness runtime. I can look into more, but first I wanted to see if this is already discussed somewhere.
Here is an incomplete list of failed tests:
/home/keceli/soft/nwx/TensorWrapper/tests/tensor/buffer/scalar_buffer.cpp:508: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/performance_tests/add.cpp:31: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/ta_helpers/einsum/einsum.cpp:56: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/diis/diis.cpp:48: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/tensor/creation.cpp:315: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/tensor/buffer/detail_/scalar_ta_buffer_pimpl.cpp:32: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/tensor/linear_algebra.cpp:90: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/tensor/linear_algebra.cpp:62: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/diis/diis.cpp:48: FAILED:
/home/keceli/soft/nwx/TensorWrapper/tests/ta_helpers/remove_redundancy.cpp:75: FAILED:
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the failures with mpirun, beginning with the listed tests such as tests/tensor/buffer/scalar_buffer.cpp and tests/tensor/creation.cpp. Compare TensorWrapper's behavior with PluginPlay and ParallelZone, then trace the MPI communicator handoff into the MADNESS runtime. Done means the affected unit tests no longer segfault when run in parallel.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- distributed-systems, testing-qa
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100