Issue 1 — branch URL in pr-link:\n\nThe new entry at perf-changelog.yaml line 4435 sets:\n\nyaml\npr-link: https://github.com/SemiAnalysisAI/InferenceX/tree/codex/minimax-m3-b300-fp4-vllm-update\n\n\nOf the ~554 pr-link values in this file, this is the only one that points to a branch tree/... URL — all others use the canonical https://github.com/SemiAnalysisAI/InferenceX/pull/<num> form (see the four most recent entries at lines 4402–4428 for examples). Since branches are typically deleted after merge, this link will 404 shortly after this PR lands. The correct value is https://github.com/SemiAnalysisAI/InferenceX/pull/1990.\n\nIssue 2 — description contradicts the diff:\n\nThe entry claims:\n\nyaml\n- "Update search space to cover more configs"\n\n\nBut the diff clearly narrows the sweep. Step-by-step count from the diff on .github/configs/nvidia-master.yaml:\n\nisl 1024, osl 1024:\n- Before (7 entries): {tp:8, c1-64}, {tp:8/ep:8, c1-512}, {tp:4, c1-64}, {tp:4/ep:4, c64-512}, {tp:4/ep:4/dp-attn, c128-512}, {tp:2/ep:2, c16-128}, {tp:8/ep:8/dp-attn, c256-1024}\n- After (4 entries): {tp:8, c1-2}, {tp:4, c2-2}, {tp:2, c4-256}, {tp:2/ep:2/dp-attn, c512-1024}\n\nisl 8192, osl 1024:\n- Before (6 entries): {tp:8, c1-64}, {tp:8/ep:8, c1-512}, {tp:4, c1-128}, {tp:4/ep:4, c64-256}, {tp:4/ep:4/dp-attn, c64-128}, {tp:8/ep:8/dp-attn, c128-256}\n- After (4 entries): {tp:8, c1-2}, {tp:4, c2-2}, {tp:2, c4-256}, {tp:2/ep:2/dp-attn, c512-1024}\n\nEvery TP8/EP8 and TP4/EP4 lane is dropped, and the TP8/TP4 lanes are drastically clipped (tp:8 from c1-64 down to c1-2; tp:4 from c1-64 down to a single c2 point). While tp:2 and tp:2/ep:2/dp-attn are genuinely new, the aggregate coverage is significantly smaller (roughly 38 concurrency points → ~12 at isl 1024). Calling this "cover more configs" is the opposite of what the change actually does.\n\nImpact:\n\nBoth issues are purely documentation quality — no runtime effect, and no CI schema check enforces the URL form. But the perf-changelog is a user-facing history document that people read to understand why a config changed; a stale link plus a misleading description defeats that purpose.\n\nFix:\n\nyaml\n- config-keys:\n - minimaxm3-fp4-b300-vllm\n description:\n - "Update Minimax M3 b300 vllm image tag"\n - "Refocus/narrow the search space and drop TP8/EP8 and TP4/EP4 sweeps."\n pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1990\n\n\nThe neighboring dsr1-fp4-b200-sglang entry at lines 4422–4428 uses similar honest phrasing when reducing coverage ("Drop the TP8/EP8 single-concurrency points.").

What the bug is

At .github/configs/nvidia-master.yaml lines 12805-12810, immediately above the minimaxm3-fp4-b300-vllm entry, there is a block comment that states:

MiniMax-M3 modelopt NVFP4 support (vllm-project/vllm PR #46380) is baked into the perf container image, so no runtime patch is needed.

That wording was true when the config used the bespoke perf-tagged image vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-8b00f41. This PR replaces that tag with vllm/vllm-openai:nightly-93d8f834dd8acf33eb0e2a75b2711b628cb6e226 — a mainline nightly image, not a perf-variant image. After the swap, the phrase "perf container image" no longer describes what is actually being pulled.

Why the invariant still holds, but the comment is misleading

The underlying claim — "no runtime patch is needed" — is still true, because vllm-project/vllm PR #46380 has presumably landed in vLLM main and the nightly image picks it up. So there is no runtime effect and nothing breaks at benchmark time. The issue is purely that the comment cites the wrong reason: a future reader inspecting this entry will look for a "perf container image" that no longer exists in the config and be confused.

Step-by-step proof

1. Before this PR, line 12812 reads image: vllm/vllm-openai:vllm-minimax-m3-perf-x86_64-13.0.1-8b00f41. The tag literally contains minimax-m3-perf, matching "perf container image".
2. This PR changes line 12812 to image: vllm/vllm-openai:nightly-93d8f834dd8acf33eb0e2a75b2711b628cb6e226. The tag is nightly-<sha>; there is no "perf" in the tag name.
3. Lines 12805-12810 (comment above the entry) still literally say baked into the perf container image, unchanged by this PR.
4. Therefore the comment and the code disagree about which image class is in use.

Fix

Update the comment to reflect that MiniMax-M3 NVFP4 support has landed upstream and is now picked up from the mainline nightly image, e.g.:

MiniMax-M3 modelopt NVFP4 support (vllm-project/vllm PR #46380) has landed in vLLM main and is picked up from the nightly image below, so no runtime patch is needed.

Scope note

The same wording appears at lines 12839-12840 above the EAGLE3 variant (minimaxm3-fp4-eagle3-b300-vllm), which does not change its image in this PR. That comment remains accurate and should be left as-is; only the copy above minimaxm3-fp4-b300-vllm (lines 12805-12810) needs updating.

Severity

This is a documentation-only inconsistency with no runtime effect on the benchmark. Marking as nit.