ISSUE3：使用GPU的CE实现deepEP-internode中跨nvlink数据转发功能
Open
Nobody has claimed this yet.
犀牛鸟-低难度
腾讯犀牛鸟开源专属
- Dominant language
- C++
- Stars
- 23
- Forks
- 38
- PR merge metrics
- No merged PRs in 30d
Description
ISSUE 目标
降低deepEP-internode场景SM使用，降低推理成本
验收要求:
- 正确运行deepEP test_internode.py。
- 通信完成时间不减少的情况下，减少跨nvlink转发使用SM/warp数50%以上
详情可见 /src/test/issue3/README.md 文件
Contributor guide
No contributing guide indexed for this repository
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 with /src/test/issue3/README.md and inspect the deepEP test_internode.py entry point. Establish the current internode communication time and SM/warp usage for cross-NVLink forwarding, then verify the implementation still passes the test while reducing forwarding SM/warp usage by more than 50% without increasing communication time.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- distributed-systems, hpc, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100