Repository navigation

refactor: migrate fixed-sequence recipes to SRT-Slurm / 将定长配方迁移至 SRT-Slurm - #3352

Merged

adibarra merged 41 commits into

mainfrom

codex/single-node-srt-slurm

Sep 24, 2026

Merged

refactor: migrate fixed-sequence recipes to SRT-Slurm / 将定长配方迁移至 SRT-Slurm#3352
adibarra merged 41 commits into
mainfrom
codex/single-node-srt-slurm

Commits on Sep 21, 2026
- refactor: start native single-node SRT-Slurm migration
  adibarra
  committed
  eafd4be
- docs: link single-node migration changelog to draft PR
  adibarra
  committed
  81f65a0
- feat: wire native H200 SRT pilot into end-to-end workflow
  adibarra
  committed
  80a4da3
- test: cover native single-node job failures and artifacts
  adibarra
  committed
  42eecdb
- fix: drop unused AIPerf inputs from SRT setup
  adibarra
  committed
  74687e3
- fix: let native SRT stage the pilot container image
  adibarra
  committed
  443ff63
- fix: bootstrap native SRT binaries before pilot submission
  adibarra
  committed
  348f773
Commits on Sep 22, 2026
- feat: port H200 MTP and Qwen recipes to native SRT
  adibarra
  committed
  a88a27a
- docs: remove migration guide
  adibarra
  committed
  b9757d9
- chore: merge main into SRT migration
  adibarra
  committed
  8762e8c
- feat: migrate fixed-sequence SGLang recipes to native SRT
  adibarra
  committed
  2763e8d
- feat: migrate fixed-sequence TRT recipes to native SRT
  adibarra
  committed
  a67435c
- chore: merge main into SRT migration
  adibarra
  committed
  b4724b0
- feat: convert remaining fixed-sequence recipes to native SRT
  adibarra
  committed
  83eaebe
- merge: sync main into native SRT migration
  adibarra
  committed
  a5e45fa
- fix: forward native Docker eval model identity
  adibarra
  committed
  c40b057
- fix: apply AMD container options as native leaf overrides
  adibarra
  committed
  571fa51
- refactor: reuse native job workspace for AMD scratch files
  adibarra
  committed
  295d7e0
- fix: complete AMD native launch and terminal status handling
  adibarra
  committed
  0bd3bf5
- refactor: require native recipes for the Slurm fixed-sequence cutover
  adibarra
  committed
  8e27918
- refactor: make single-node fixed-sequence coverage SRT-only
  adibarra
  committed
  8f3f0d5
- fix: allow client dependencies in container virtual environments
  adibarra
  committed
  c430090
- fix: limit native single-node steps to serving GPUs
  adibarra
  committed
  6f60a8b
- chore: merge main into single-node SRT migration
  adibarra
  committed
  b3a5a7f
Commits on Sep 23, 2026
- fix: restore repository workdir for native container steps
  adibarra
  committed
  161aa24
- chore: sync latest main before ATOM smoke retry
  adibarra
  committed
  838607d
- fix: stream AMD power samples without input buffering
  adibarra
  committed
  dd95570
- chore: sync main into SRT migration
  adibarra
  committed
  2d532e6
- fix: match registered H200 runner labels
  adibarra
  committed
  77961ed
- chore: merge main into SRT migration
  adibarra
  committed
  814c007
- chore: sync GB300 AgentX updates from main
  adibarra
  committed
  c6f47fa
- chore: sync H200 AgentX updates from main
  adibarra
  committed
  4747349
- chore: pin srt-slurm submodule to main [skip ci]
  cquil11
  committed
  a3dc898
- refactor(srt): return single-node migration to NVIDIA upstream
  adibarra
  committed
  082279d
- merge: preserve concurrent upstream submodule update
  adibarra
  committed
  66781f6
- merge: sync main result-processing Python fix
  adibarra
  committed
  2e358d1
- chore: bump srt-slurm submodule to v2.23.2
  cquil11
  committed
  b0a9064
Commits on Sep 24, 2026

Commits

Commits on Sep 21, 2026

Commits on Sep 22, 2026

Commits on Sep 23, 2026

Commits on Sep 24, 2026