Skip to content

Commit a9339df

Browse files
committed
chore(perf-changelog): trigger multinode sweep for measured-power aggregation
Appends entry for dsv4-fp4-gb300-dynamo-sglang so run-sweep.yml fires when the sweep-enabled label is added to PR #1574. The sweep produces the first multinode agg JSONs with avg_power_w + joules_per_*_token, validating the per-source GPU-id namespacing and GPU_METRICS_CSV_GLOB env-var bridge end-to-end on real GB300 hardware (gb300-cw cluster).
1 parent 5e6533d commit a9339df

1 file changed

Lines changed: 6 additions & 0 deletions

File tree

perf-changelog.yaml

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3192,3 +3192,9 @@
31923192
- "Add GLM-5-FP8 models.yaml flags, setup_deps.sh (aiter gluon + transformers glm_moe_dsa), GLM-5 env tuning in env.sh"
31933193
- "Add multinode launch script glm5_fp8_mi355x_sglang-disagg.sh; server.sh sources setup_deps.sh"
31943194
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1572
3195+
3196+
- config-keys:
3197+
- dsv4-fp4-gb300-dynamo-sglang
3198+
description:
3199+
- "Smoke run validating multinode measured-power aggregation (PR #1574). No config change; entry exists to trigger a sweep that produces the first multinode agg JSON with avg_power_w + joules_per_*_token populated from per-node srt-slurm perfmon CSVs. Validates per-source GPU-id namespacing in aggregate_power.py (without it, 14 nodes × 4 GPUs would report num_gpus=4 instead of 56) and the GPU_METRICS_CSV_GLOB env var bridge in process_result.py. Only the gb300-cw runner has the perfmon launcher changes; any gb300-nv runs in the sweep will succeed normally without power fields, which the dashboard handles gracefully (chart gates on field presence)."
3200+
pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1574

0 commit comments

Comments
 (0)