deepspeedai / deepspeedai/DeepSpeed
so many cudafree api called between nccl allgather when larger batchsize and zero2 config is enabled during training
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 43.1k
- Forks
- 5k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 112
Description
This issue has no description.
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
The issue names no file, test, or entry point. Start by locating the NCCL all-gather path used with ZeRO-2 and larger batch sizes; done means explaining and addressing the unexpected CUDA-free API calls, with a regression check for that configuration.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- distributed-systems, performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100