[NV] minimaxm3-fp4-b200-dynamo-vllm-mtp

benchmarks/multi_node/srt-slurm-recipes/vllm/minimax-m3/b200-fp4/8k1k/mtp/1p1d-dep4-tp4-eagle3-8k1k.yaml

@@ -0,0 +1,94 @@
+name: "minimax-m3-vllm-disagg-b200-1p1d-dep4-tp4-fp4-8k1k-eagle3"
+
+model:
+  path: "nvidia/MiniMax-M3-NVFP4"
+  container: "vllm/vllm-openai:nightly-6f91edf96d3f3272945809c04702380053bff4de"
+  precision: fp4
+
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  prefill_nodes: 1
+  decode_nodes: 0
+  prefill_workers: 1
+  decode_workers: 1
+  gpus_per_prefill: 4
+  gpus_per_decode: 4
+
+dynamo:
+  install: true
+  version: 1.3.0.dev20260710
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+
+backend:
+  type: vllm
+  connector: null
+  allow_prefill_decode_colocation: true
+  allow_prefill_decode_colocation_across_nodes: true
+
+  prefill_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  decode_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  vllm_config:
+    prefill:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 1
+      pipeline-parallel-size: 1
+      data-parallel-size: 4
+      data-parallel-rpc-port: 13345
+      enable-expert-parallel: true
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-cudagraph-capture-size: 2048
+      max-num-batched-tokens: 16384
+
+    decode:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 4
+      pipeline-parallel-size: 1
+      enable-expert-parallel: false
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-num-seqs: 1024
+      max-num-batched-tokens: 16384
+      max-cudagraph-capture-size: 2048
+
+health_check:
+  max_attempts: 360
+  interval_seconds: 10
+
+benchmark:
+  type: sa-bench
+  isl: 8192
+  osl: 1024
+  req_rate: inf
+  num_warmup_mult: 0
+  random_range_ratio: 0.8
+  use_chat_template: true
+  concurrencies: "1x2x4x8x16x32x64x128x256x512x1024"

benchmarks/multi_node/srt-slurm-recipes/vllm/minimax-m3/b200-fp4/8k1k/mtp/1p1d-dep4-tp8-eagle3-8k1k.yaml

@@ -0,0 +1,94 @@
+name: "minimax-m3-vllm-disagg-b200-1p1d-dep4-tp8-fp4-8k1k-eagle3"
+
+model:
+  path: "nvidia/MiniMax-M3-NVFP4"
+  container: "vllm/vllm-openai:nightly-6f91edf96d3f3272945809c04702380053bff4de"
+  precision: fp4
+
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  prefill_nodes: 1
+  decode_nodes: 1
+  prefill_workers: 1
+  decode_workers: 1
+  gpus_per_prefill: 4
+  gpus_per_decode: 8
+
+dynamo:
+  install: true
+  version: 1.3.0.dev20260710
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+
+backend:
+  type: vllm
+  connector: null
+  allow_prefill_decode_colocation: true
+  allow_prefill_decode_colocation_across_nodes: true
+
+  prefill_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  decode_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  vllm_config:
+    prefill:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 1
+      pipeline-parallel-size: 1
+      data-parallel-size: 4
+      data-parallel-rpc-port: 13345
+      enable-expert-parallel: true
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-cudagraph-capture-size: 2048
+      max-num-batched-tokens: 16384
+
+    decode:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 8
+      pipeline-parallel-size: 1
+      enable-expert-parallel: false
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-num-seqs: 1024
+      max-num-batched-tokens: 16384
+      max-cudagraph-capture-size: 2048
+
+health_check:
+  max_attempts: 360
+  interval_seconds: 10
+
+benchmark:
+  type: sa-bench
+  isl: 8192
+  osl: 1024
+  req_rate: inf
+  num_warmup_mult: 0
+  random_range_ratio: 0.8
+  use_chat_template: true
+  concurrencies: "1x2x4x8x16x32x64x128x256x512x1024"

benchmarks/multi_node/srt-slurm-recipes/vllm/minimax-m3/b200-fp4/8k1k/mtp/1p2d-dep4-tp4-eagle3-8k1k.yaml

@@ -0,0 +1,94 @@
+name: "minimax-m3-vllm-disagg-b200-1p2d-dep4-tp4-fp4-8k1k-eagle3"
+
+model:
+  path: "nvidia/MiniMax-M3-NVFP4"
+  container: "vllm/vllm-openai:nightly-6f91edf96d3f3272945809c04702380053bff4de"
+  precision: fp4
+
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  prefill_nodes: 1
+  decode_nodes: 1
+  prefill_workers: 1
+  decode_workers: 2
+  gpus_per_prefill: 4
+  gpus_per_decode: 4
+
+dynamo:
+  install: true
+  version: 1.3.0.dev20260710
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+
+backend:
+  type: vllm
+  connector: null
+  allow_prefill_decode_colocation: true
+  allow_prefill_decode_colocation_across_nodes: true
+
+  prefill_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  decode_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  vllm_config:
+    prefill:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 1
+      pipeline-parallel-size: 1
+      data-parallel-size: 4
+      data-parallel-rpc-port: 13345
+      enable-expert-parallel: true
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-cudagraph-capture-size: 2048
+      max-num-batched-tokens: 16384
+
+    decode:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 4
+      pipeline-parallel-size: 1
+      enable-expert-parallel: false
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-num-seqs: 1024
+      max-num-batched-tokens: 16384
+      max-cudagraph-capture-size: 2048
+
+health_check:
+  max_attempts: 360
+  interval_seconds: 10
+
+benchmark:
+  type: sa-bench
+  isl: 8192
+  osl: 1024
+  req_rate: inf
+  num_warmup_mult: 0
+  random_range_ratio: 0.8
+  use_chat_template: true
+  concurrencies: "1x2x4x8x16x32x64x128x256x512x1024"

benchmarks/multi_node/srt-slurm-recipes/vllm/minimax-m3/b200-fp4/8k1k/mtp/1p2d-dep4-tp8-eagle3-8k1k.yaml

@@ -0,0 +1,94 @@
+name: "minimax-m3-vllm-disagg-b200-1p2d-dep4-tp8-fp4-8k1k-eagle3"
+
+model:
+  path: "nvidia/MiniMax-M3-NVFP4"
+  container: "vllm/vllm-openai:nightly-6f91edf96d3f3272945809c04702380053bff4de"
+  precision: fp4
+
+resources:
+  gpu_type: b200
+  gpus_per_node: 8
+  prefill_nodes: 1
+  decode_nodes: 2
+  prefill_workers: 1
+  decode_workers: 2
+  gpus_per_prefill: 4
+  gpus_per_decode: 8
+
+dynamo:
+  install: true
+  version: 1.3.0.dev20260710
+
+frontend:
+  type: dynamo
+  enable_multiple_frontends: false
+
+backend:
+  type: vllm
+  connector: null
+  allow_prefill_decode_colocation: true
+  allow_prefill_decode_colocation_across_nodes: true
+
+  prefill_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  decode_environment:
+    VLLM_FLOAT32_MATMUL_PRECISION: high
+    VLLM_FLASHINFER_ALLREDUCE_BACKEND: trtllm
+    UCX_TLS: cuda_copy,cuda_ipc,rc
+
+  vllm_config:
+    prefill:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 1
+      pipeline-parallel-size: 1
+      data-parallel-size: 4
+      data-parallel-rpc-port: 13345
+      enable-expert-parallel: true
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-cudagraph-capture-size: 2048
+      max-num-batched-tokens: 16384
+
+    decode:
+      no-enable-flashinfer-autotune: true
+      tensor-parallel-size: 8
+      pipeline-parallel-size: 1
+      enable-expert-parallel: false
+      trust-remote-code: true
+      no-enable-prefix-caching: true
+      kv-transfer-config: '{"kv_connector": "NixlConnector", "kv_role": "kv_both"}'
+      attention-config: '{"backend": "FLASHINFER", "use_trtllm_attention": true, "indexer_kv_dtype": "fp8"}'
+      block-size: 128
+      gpu-memory-utilization: 0.95
+      max-model-len: 9472
+      language-model-only: true
+      speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASH_ATTN"}'
+      stream-interval: 32
+      max-num-seqs: 1024
+      max-num-batched-tokens: 16384
+      max-cudagraph-capture-size: 2048
+
+health_check:
+  max_attempts: 360
+  interval_seconds: 10
+
+benchmark:
+  type: sa-bench
+  isl: 8192
+  osl: 1024
+  req_rate: inf
+  num_warmup_mult: 0
+  random_range_ratio: 0.8
+  use_chat_template: true
+  concurrencies: "1x2x4x8x16x32x64x128x256x512x1024"