deepseek-ai / deepseek-ai/DeepEP
output tokens in intranode::dispatch for each expert isn't packed?
Open
- Dominant language
- Cuda
- Stars
- 10.1k
- Forks
- 1.4k
- Avg merge
- 4d 1h
- Merged PRs (30d)
- 2
Description
The output token in the intranode::dispatch kernel offsets is channel_offset + rank_offsets + recv_token_idx;
The tokens for each expert isn't contiguous?
If yes, do you have plan to optimize these?
Thanks.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.