From 08ed9f3e861760fde9773f1124e3bbb6497277d5 Mon Sep 17 00:00:00 2001
From: functionstackx <47992694+functionstackx@users.noreply.github.com>
Date: Thu, 25 Jun 2026 15:27:10 -0400
Subject: [PATCH 1/2] feat: add MiniMax M3 FP4 MI355X vLLM benchmark

---
 .github/configs/amd-master.yaml               | 30 +++++++
 .../minimaxm3_fp4_mi355x_vllm.sh              | 87 +++++++++++++++++++
 perf-changelog.yaml                           |  9 ++
 3 files changed, 126 insertions(+)
 create mode 100755 benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh

diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml
index 6e06029ff..bafc37688 100644
--- a/.github/configs/amd-master.yaml
+++ b/.github/configs/amd-master.yaml
@@ -2636,6 +2636,36 @@ minimaxm3-fp4-mi355x-vllm-disagg:
           dp-attn: false
           additional-settings:
           - "DECODE_NODES=1"
+# MiniMax-M3 MXFP4 MI355X vLLM recipe. The pinned nightly includes upstream
+# MiniMax-M3 Quark MXFP4 support (vllm-project/vllm#45794). Use the text-only
+# language-model path and mirror the MXFP8 MI355X search space for a direct
+# precision comparison.
+minimaxm3-fp4-mi355x-vllm:
+  image: vllm/vllm-openai-rocm:nightly-3f5a1e1733200760169ff31ebe60a271072b199e
+  model: amd/MiniMax-M3-MXFP4
+  model-prefix: minimaxm3
+  runner: mi355x
+  precision: fp4
+  framework: vllm
+  multinode: false
+  scenarios:
+    fixed-seq-len:
+    - isl: 1024
+      osl: 1024
+      search-space:
+      - { tp: 8, conc-start: 1, conc-end: 64 }
+      - { tp: 8, ep: 8, conc-start: 1, conc-end: 512 }
+      - { tp: 4, conc-start: 1, conc-end: 64 }
+      - { tp: 4, ep: 4, conc-start: 64, conc-end: 512 }
+      - { tp: 2, ep: 2, conc-start: 16, conc-end: 128 }
+      - { tp: 8, ep: 8, dp-attn: true, conc-start: 256, conc-end: 1024 }
+    - isl: 8192
+      osl: 1024
+      search-space:
+      - { tp: 8, conc-start: 1, conc-end: 64 }
+      - { tp: 8, ep: 8, conc-start: 1, conc-end: 512 }
+      - { tp: 4, conc-start: 1, conc-end: 128 }
+      - { tp: 8, ep: 8, dp-attn: true, conc-start: 128, conc-end: 512 }
 
 # MiniMax-M3 MXFP4 MI355X atom recipe:
 # https://github.com/ROCm/ATOM/blob/5d42d49f9e4292e5b61475917e92e7ec1b1dacb7/recipes/MiniMax-M3.md
diff --git a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh
new file mode 100755
index 000000000..52ad82fd7
--- /dev/null
+++ b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh
@@ -0,0 +1,87 @@
+#!/usr/bin/env bash
+
+# MiniMax-M3 MXFP4 MI355X (gfx950) single-node vLLM recipe.
+# https://huggingface.co/amd/MiniMax-M3-MXFP4#reproduction
+# Block size 128 is mandatory for MSA. This fixed-sequence benchmark uses the
+# text-only language-model path and lets vLLM select the MoE backend.
+
+source "$(dirname "$0")/../../benchmark_lib.sh"
+
+check_env_vars \
+    MODEL \
+    TP \
+    EP_SIZE \
+    DP_ATTENTION \
+    CONC \
+    ISL \
+    OSL \
+    MAX_MODEL_LEN \
+    RANDOM_RANGE_RATIO \
+    RESULT_FILENAME
+
+if [[ -n "$SLURM_JOB_ID" ]]; then
+  echo "JOB $SLURM_JOB_ID running on $SLURMD_NODENAME"
+fi
+
+if [[ "$MODEL" != /* ]]; then hf download "$MODEL"; fi
+
+if [ -n "$ROCR_VISIBLE_DEVICES" ]; then
+    export HIP_VISIBLE_DEVICES="$ROCR_VISIBLE_DEVICES"
+fi
+
+SERVER_LOG=/workspace/server.log
+export VLLM_ENGINE_READY_TIMEOUT_S=3600
+
+if [ "${EVAL_ONLY}" = "true" ]; then
+    setup_eval_context
+fi
+
+PARALLEL_ARGS=(--tensor-parallel-size "$TP")
+if [ "${DP_ATTENTION}" = "true" ]; then
+    PARALLEL_ARGS=(
+        --tensor-parallel-size 1
+        --data-parallel-size "$TP"
+        --enable-expert-parallel
+    )
+elif [ "$EP_SIZE" -gt 1 ]; then
+    PARALLEL_ARGS+=(--enable-expert-parallel)
+fi
+
+start_gpu_monitor
+
+set -x
+vllm serve "$MODEL" --port "$PORT" \
+    "${PARALLEL_ARGS[@]}" \
+    --trust-remote-code \
+    --block-size 128 \
+    --no-enable-prefix-caching \
+    --language-model-only \
+    --max-model-len "$MAX_MODEL_LEN" \
+    --attention-backend TRITON_ATTN \
+    --tool-call-parser minimax_m3 \
+    --enable-auto-tool-choice \
+    --reasoning-parser minimax_m3 > "$SERVER_LOG" 2>&1 &
+
+SERVER_PID=$!
+wait_for_server_ready --port "$PORT" --server-log "$SERVER_LOG" --server-pid "$SERVER_PID"
+
+run_benchmark_serving \
+    --model "$MODEL" \
+    --port "$PORT" \
+    --backend vllm \
+    --input-len "$ISL" \
+    --output-len "$OSL" \
+    --random-range-ratio "$RANDOM_RANGE_RATIO" \
+    --num-prompts "$((CONC * 10))" \
+    --max-concurrency "$CONC" \
+    --result-filename "$RESULT_FILENAME" \
+    --result-dir /workspace/ \
+    --trust-remote-code
+
+if [ "${RUN_EVAL}" = "true" ]; then
+    run_eval --framework lm-eval --port "$PORT"
+    append_lm_eval_summary
+fi
+
+stop_gpu_monitor
+set +x
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index f504fc4d9..81009f2c4 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4196,3 +4196,12 @@
   description:
     - "Initial submission: MiniMax-M3 MXFP4 disagg (prefill/decode) on MI355X with vLLM over the MoRI-IO KV connector (8k/1k)."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1914
+
+- config-keys:
+    - minimaxm3-fp4-mi355x-vllm
+  description:
+    - "Add a MiniMax-M3 MXFP4 single-node vLLM benchmark on MI355X using amd/MiniMax-M3-MXFP4."
+    - "Pin vllm/vllm-openai-rocm:nightly-3f5a1e1733200760169ff31ebe60a271072b199e, which includes upstream Quark MXFP4 support from vllm-project/vllm#45794."
+    - "Serve through the text-only language-model path with block size 128, TRITON_ATTN, MiniMax-M3 tool/reasoning parsers, and automatic tool choice; let vLLM select the MoE backend and retain the default KV-cache dtype."
+    - "Mirror the MiniMax-M3 MXFP8 MI355X TP/EP/DP-attention search space at 1k1k and 8k1k for a direct precision comparison."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1935

From 87cf65e97fea77f2b7d91fa0f19ed87a7dcc71d5 Mon Sep 17 00:00:00 2001
From: functionstackx <47992694+functionstackx@users.noreply.github.com>
Date: Thu, 25 Jun 2026 16:31:39 -0400
Subject: [PATCH 2/2] fix: disable breakable cudagraph for MiniMax M3 FP4

---
 .../single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh      | 1 +
 perf-changelog.yaml                                             | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh
index 52ad82fd7..16ec09e77 100755
--- a/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh
+++ b/benchmarks/single_node/fixed_seq_len/minimaxm3_fp4_mi355x_vllm.sh
@@ -31,6 +31,7 @@ fi
 
 SERVER_LOG=/workspace/server.log
 export VLLM_ENGINE_READY_TIMEOUT_S=3600
+export VLLM_USE_BREAKABLE_CUDAGRAPH=0
 
 if [ "${EVAL_ONLY}" = "true" ]; then
     setup_eval_context
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 81009f2c4..7c16dc460 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -4202,6 +4202,6 @@
   description:
     - "Add a MiniMax-M3 MXFP4 single-node vLLM benchmark on MI355X using amd/MiniMax-M3-MXFP4."
     - "Pin vllm/vllm-openai-rocm:nightly-3f5a1e1733200760169ff31ebe60a271072b199e, which includes upstream Quark MXFP4 support from vllm-project/vllm#45794."
-    - "Serve through the text-only language-model path with block size 128, TRITON_ATTN, MiniMax-M3 tool/reasoning parsers, and automatic tool choice; let vLLM select the MoE backend and retain the default KV-cache dtype."
+    - "Serve through the text-only language-model path with block size 128, TRITON_ATTN, MiniMax-M3 tool/reasoning parsers, automatic tool choice, and VLLM_USE_BREAKABLE_CUDAGRAPH=0; let vLLM select the MoE backend and retain the default KV-cache dtype."
     - "Mirror the MiniMax-M3 MXFP8 MI355X TP/EP/DP-attention search space at 1k1k and 8k1k for a direct precision comparison."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1935