Skip to content

Expert distribution recording without overhead for EPLB#4957

Merged
zhyncs merged 285 commits intosgl-project:mainfrom
fzyzcjy:feat/expert_distribution_recorder
May 20, 2025
Merged

Expert distribution recording without overhead for EPLB#4957
zhyncs merged 285 commits intosgl-project:mainfrom
fzyzcjy:feat/expert_distribution_recorder

Conversation

@fzyzcjy
Copy link
Copy Markdown
Collaborator

@fzyzcjy fzyzcjy commented Apr 1, 2025

Motivation

For EPLB, and also for debugging/knowing details

dep: #5219

NOTE: There are enhancements to this, but it currently in branch #5295 and not yet extracted to here.

Modifications

Checklist

@fzyzcjy
Copy link
Copy Markdown
Collaborator Author

fzyzcjy commented May 20, 2025

➜  misc (cd /host_home/primary_synced/sglang && SGL_DISABLE_TP_MEMORY_INBALANCE_CHECK=1 python3 test/srt/test_full_deepseek_v3.py)
➜  misc (cd /host_home/primary_synced/sglang && SGL_DISABLE_TP_MEMORY_INBALANCE_CHECK=1 python3 test/srt/test_disaggregation_different_tp.py)

image

image

@zhyncs zhyncs merged commit f065388 into sgl-project:main May 20, 2025
1 of 42 checks passed
Layssy pushed a commit to Layssy/sglang-iaas that referenced this pull request Jun 9, 2025
xwu-intel pushed a commit to xwu-intel/sglang that referenced this pull request Jun 17, 2025
Yuechguo pushed a commit to Yuechguo/sglang that referenced this pull request Jul 29, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants