[AgentX] add Muse Spark Open NVFP4 on B300 / [AgentX] 新增 B300 Muse Spark Open NVFP4 支持

inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/musespark/sglang/b300-fp4-mtp/agentic.yaml

@@ -0,0 +1,68 @@
+base:
+  schema: 2
+  name: musespark-fp4-b300-sglang-agentic
+  resources:
+    gpu_type: b300
+    gpus_per_node: 8
+  slurm:
+    time_limit: '4:00:00'
+  frontend:
+    type: sglang
+    enable_multiple_frontends: false
+  observability:
+    enabled: false
+    tachometer:
+      enabled: false
+  engine: sglang
+  roles:
+    agg:
+      nodes: 1
+      workers: 1
+      gpus: 4
+      args:
+        served-model-name: meta-models/Muse-Spark-Open-816B-A42B-NVFP4
+        trust-remote-code: true
+        tensor-parallel-size: 4
+        expert-parallel-size: 4
+        quantization: modelopt_fp4
+        moe-runner-backend: flashinfer_trtllm
+        kv-cache-dtype: fp8_e4m3
+        mem-fraction-static: 0.8
+        enable-layernorm-sp: true
+        enable-megamoe-for-layernorm-sp: true
+        # DFlash2 verifies 8 tokens per step: the current token plus 7 drafts.
+        speculative-algorithm: DFLASH
+        speculative-draft-model-path: RadixArk/Muse-Spark-Open-DFlash2
+        speculative-draft-attention-backend: fa4
+        speculative-num-draft-tokens: 8
+        mm-feature-transport: cpu
+        reasoning-parser: muse
+        tool-call-parser: muse
+        watchdog-timeout: 3600
+        enable-metrics: true
+      env:
+        PYTHONNOUSERSITE: '1'
+        PYTHONUNBUFFERED: '1'
+        # Outlast AIPerf's pooled connections past Uvicorn's keep-alive.
+        SGLANG_TIMEOUT_KEEP_ALIVE: '900'
+  # Cold 533 GB weight loads from shared storage plus graph capture outlast the default.
+  health_check:
+    max_attempts: 1440
+    interval_seconds: 5
+  benchmark:
+    env:
+      # The gated Hub repo is unreadable with the job token; tokenize from the staged checkpoint.
+      TOKENIZER: /model
+      # Warmup can leave bytes unacknowledged past AIPerf's 30 s default.
+      AIPERF_HTTP_TCP_USER_TIMEOUT: '900000'
+      AIPERF_REQUIRED_SERVER_METRIC_PREFIX: 'sglang:'
+
+zip_override_tp4:
+  roles:
+    agg:
+      args:
+        max-running-requests: [2, 4, 8, 16, 32, 48, 64, 96, 128, 256]
+        cuda-graph-max-bs-decode: [2, 4, 8, 16, 32, 48, 64, 96, 128, 256]
+  benchmark:
+    env:
+      CONC: ['1', '2', '4', '8', '16', '24', '32', '48', '64', '128']

inferencex-e2e/configs/nvidia-master.yaml

@@ -9690,3 +9690,17 @@ dsv4-fp4-gb200-dynamo-vllm-agentic-dspark-disagg:
           tp: 16
           ep: 16
           dp-attn: true
+
+musespark-fp4-b300-sglang-agentic-dflash:
+  image: lmsysorg/sglang:dev-muse-spark-open@sha256:8cfc5a72cb296771f5f8a4a4688cb6d05685610a830cbb616fa4ab0eb7bb2b0a
+  model: meta-models/Muse-Spark-Open-816B-A42B-NVFP4
+  model-prefix: musespark
+  runner: cluster:b300-dsxe
+  precision: fp4
+  framework: sglang
+  multinode: false
+  srt-recipe-dir: musespark/sglang/b300-fp4-mtp
+  scenarios:
+    agentic-coding:
+    - search-space:
+      - { tp: 4, ep: 4, kv-offloading: none, spec-decoding: draft_model, conc-list: [1, 2, 4, 8, 16, 24, 32, 48, 64, 128], srt-recipe: agentic.yaml }

inferencex-e2e/configs/runners.yaml

@@ -525,6 +525,7 @@ clusters:
         MiniMax-M3: {root: scratch, dir: MiniMax-M3}
         MiniMax-M3-MXFP8: {root: scratch, dir: MiniMax-M3-MXFP8}
         MiniMax-M3-NVFP4: {root: scratch, dir: MiniMax-M3-NVFP4}
+        Muse-Spark-Open-816B-A42B-NVFP4: {root: data-models, dir: Muse-Spark-Open-816B-A42B-NVFP4}
         Qwen3.5-397B-A17B-FP8: {root: scratch, dir: Qwen3.5-397B-A17B-FP8}
         Qwen3.5-397B-A17B-NVFP4: {root: scratch, dir: Qwen3.5-397B-A17B-NVFP4}
         Qwen3.5-397B-A17B-NVFP4-V2: {root: scratch, dir: Qwen3.5-397B-A17B-NVFP4-V2}

inferencex-e2e/perf-changelog.yaml

@@ -9913,3 +9913,9 @@
     - "Synthetic acceptance follows the committed musespark_dflash golden curve: AL 4.05 at 7 draft tokens, thinking_on."
     - "The checkpoint is served from the staged Lustre copy; the AgentX client reads its tokenizer from it."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4140
+
+- config-keys: [musespark-fp4-b300-sglang-agentic-dflash]
+  description:
+    - "Add Muse Spark Open NVFP4 SGLang DFlash2 AgentX on B300 at TP4/EP4 with GPU-resident KV and concurrency 1 to 128."
+    - "Use the committed golden curve for DFlash acceptance and the staged checkpoint for serving and tokenization."
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4146