[RFE]: infiniband counters are needed to help diagnose the issue
Open
Nobody has claimed this yet.
enhancement
- Dominant language
- C++
- Stars
- 586
- Forks
- 110
- PR merge metrics
- No merged PRs in 30d
Description
Please provide the below details to ensure we understand your needs
Hello,
It seems that infiniband counters documented in https://enterprise-support.nvidia.com/s/article/understanding-mlx5-linux-counters-and-status-parameters are not visible, such as OOS. We think it's due to the usage of th devx.
Not sure it's the right place to ask for the support. But this has indeed become an obstacle for us when using NVSHMEM.
Thanks
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
Review how NVSHMEM uses DEVX and compare the counters it exposes with the linked NVIDIA documentation, focusing on the missing OOS counter. Done means the required InfiniBand counters are visible and usable for diagnosing NVSHMEM issues.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- distributed-systems, networking
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100