ISSUE8：基于IBGDA和IBRC的Alltoall在各个msg的理论性能上限
Open
Nobody has claimed this yet.
犀牛鸟-中高难度
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
探究GPU通信各个原子模块的耗时
验收要求:
- 输出alltoall各个msg下，ibrc和ibgda的理论耗时
- 输出alltoall的微观实现中，各个独立子模块的理论耗时计算公式
详情可见 /src/test/issue8/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 by reading /src/test/issue8/README.md to understand the Alltoall micro-implementation and the IBRC and IBGDA modules. Derive the theoretical time formulas for each independent submodule, then produce the theoretical Alltoall times for each message size and both communication methods. Done means the requested formulas and results are documented.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- distributed-systems, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100