deepseek-ai / deepseek-ai/DeepEP
Can DeepEP run correctly in cudaGraph mode?
Open
- Dominant language
- Cuda
- Stars
- 10.1k
- Forks
- 1.4k
- Avg merge
- 4d 1h
- Merged PRs (30d)
- 2
Description
Two questions:
1. Can DeepEP run normally in cudaGraph mode?
2. Does DeepEP perform "dropless" MoE dispatch? (i.e. no token discarded if tokens are heavily routed to a limited number of experts)
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by reviewing the repository's cudaGraph-related entry points and MoE dispatch documentation, if present. Determine whether DeepEP supports cudaGraph execution and whether its dispatch is dropless, then document the evidence or provide a minimal reproducible result.
Written by the indexing model from the issue text.
Assessment
- Domain
- distributed-systems, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100