[GH-ISSUE #302] test _internode with 2 nodes(H800) failed: NCCL Connect to 192.168.0.157<47713> failed : Software caused connection abort #216

New issue

Closed

opened 2026-10-26 09:56:45 +00:00 by bsakaguchi · 2 comments

bsakaguchi commented

2026-10-26 09:56:45 +00:00

Owner

Originally created by @KrisLu999 on GitHub (Jul 14, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/302

Does anyone know how to r

esolve this issue? The log is as follows:

Originally created by @KrisLu999 on GitHub (Jul 14, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/302 Does anyone know how to r <img width="482" height="370" alt="Image" src="https://github.com/user-attachments/assets/8737bf1e-8e8b-4195-bb7a-ebbf4eb8d893" /> esolve this issue? The log is as follows:

bsakaguchi closed this issue

2026-10-26 09:56:46 +00:00

bsakaguchi commented

2026-10-26 09:56:47 +00:00

Author

Owner

@sphish commented on GitHub (Jul 15, 2025):

This appears to be an error encountered during NCCL initialization. You can extract a minimal reproducible script and then ask in the NCCL repo.

@sphish commented on GitHub (Jul 15, 2025): This appears to be an error encountered during NCCL initialization. You can extract a minimal reproducible script and then ask in the NCCL repo.

bsakaguchi commented

2026-10-26 09:56:48 +00:00

Author

Owner

@KrisLu999 commented on GitHub (Jul 29, 2025):

This appears to be an error encountered during NCCL initialization. You can extract a minimal reproducible script and then ask in the NCCL repo.

thank you

@KrisLu999 commented on GitHub (Jul 29, 2025): > This appears to be an error encountered during NCCL initialization. You can extract a minimal reproducible script and then ask in the NCCL repo. thank you

bsakaguchi referenced this issue

2026-10-26 09:56:49 +00:00

[GH-ISSUE #307] Test test_low_latency.py on L20*8 failed #217