[GH-ISSUE #25] NVlink latency #26

New issue

Closed

opened 2026-10-26 09:52:43 +00:00 by bsakaguchi · 2 comments

bsakaguchi commented

2026-10-26 09:52:43 +00:00

Owner

Originally created by @zhang662817 on GitHub (Feb 26, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/25

In Intranode, pure rdma latency is better than nvlink? could you show latency comparison data between nvlink and rdma.

Originally created by @zhang662817 on GitHub (Feb 26, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/25 In Intranode, pure rdma latency is better than nvlink? could you show latency comparison data between nvlink and rdma.

bsakaguchi closed this issue

2026-10-26 09:52:44 +00:00

bsakaguchi commented

2026-10-26 09:52:45 +00:00

Author

Owner

@sphish commented on GitHub (Feb 26, 2025):

As mentioned in https://github.com/deepseek-ai/DeepEP/issues/22#issuecomment-2684270222, our low latency kernels are optimized for internode communication. We have not yet implemented pure NVLink low-latency kernels.

@sphish commented on GitHub (Feb 26, 2025): As mentioned in https://github.com/deepseek-ai/DeepEP/issues/22#issuecomment-2684270222, our low latency kernels are optimized for internode communication. We have not yet implemented pure NVLink low-latency kernels.

bsakaguchi commented

2026-10-26 09:52:46 +00:00

Author

Owner

@zhang662817 commented on GitHub (Feb 27, 2025):

Thanks

@zhang662817 commented on GitHub (Feb 27, 2025): Thanks

bsakaguchi referenced this issue

2026-10-26 09:53:39 +00:00

[GH-ISSUE #74] test_internode.py hangs at internode.cu:dispatch(...) #68

bsakaguchi referenced this issue

2026-10-26 09:56:29 +00:00