Skip to content

config(minimaxm3): update GB200 Dynamo vLLM AgentX recipes / 更新 GB200 Dynamo vLLM AgentX 配方 #3424

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
RohitNagraj wants to merge 2 commits into
base: main
Choose a base branch
from
Open

config(minimaxm3): update GB200 Dynamo vLLM AgentX recipes / 更新 GB200 Dynamo vLLM AgentX 配方 #3424

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b87dc97
refactor(minimaxm3): align GB200 recipes with e2e layout
RohitNagraj Sep 28, 2026
0a90206
fix(minimaxm3): remove GB200 disaggregated C50 point
RohitNagraj Sep 28, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
109 changes: 109 additions & 0 deletions ...urm-recipes/minimaxm3/vllm/gb200-fp4/agentx/agg-tp4-vllm-simple-mns16-nightly-native.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,109 @@
schema: 2
name: "minimax-m3-vllm-agg-gb200-tp4-vllm-simple-mns16-agentic-nightly-native"

model:
path: "minimax-m3-nvfp4"
container: "vllm/vllm-openai:nightly-9ea8f3ffc354901b740f0b31988900897b7221d7"
precision: "fp4"

identity:
model: {repo: "nvidia/MiniMax-M3-NVFP4"}
container: {image: "vllm/vllm-openai:nightly-9ea8f3ffc354901b740f0b31988900897b7221d7"}
frameworks: {dynamo: "1.5.0.dev20260908"}

dynamo: {install: true, source: {pypi: "1.5.0.dev20260908"}}
environment: {ETCD_LEASE_TTL: "7200"}
slurm: {time_limit: "12:00:00"}
health_check: {max_attempts: 2160, interval_seconds: 10}
resources: {gpu_type: "gb200", gpus_per_node: 4}
services:
- name: etcd
type: etcd
placement:
node: infra
- name: nats
type: nats
placement:
node: infra
options:
max_payload_mb: 32

frontend:
type: dynamo
enable_multiple_frontends: false
env: {DYN_TCP_REQUEST_TIMEOUT: "60"}
args:
trust-remote-code: true
router-mode: "kv"
router-kv-events: true
router-temperature: "0"
router-session-affinity-ttl-secs: 14400
kv-cache-block-size: 128

engine:
type: vllm
connector:
roles:
agg:
nodes: 1
workers: 1
gpus: 4
env:
VLLM_ENGINE_READY_TIMEOUT_S: "7200"
VLLM_EXECUTE_MODEL_TIMEOUT_SECONDS: "1800"
VLLM_FLOAT32_MATMUL_PRECISION: "high"
VLLM_FLASHINFER_ALLREDUCE_BACKEND: "trtllm"
VLLM_LOG_STATS_INTERVAL: "1"
VLLM_USE_SIMPLE_KV_OFFLOAD: "1"
NCCL_CUMEM_ENABLE: "1"
NCCL_MNNVL_ENABLE: "1"
NCCL_NVLS_ENABLE: "1"
UCX_MEMTYPE_CACHE: "n"
UCX_NET_DEVICES: "mlx5_0:1,mlx5_1:1,mlx5_2:1,mlx5_3:1"
UCX_TLS: "cuda_copy,cuda_ipc,rc"
NCCL_IB_HCA: "mlx5_0,mlx5_1,mlx5_2,mlx5_3"
args:
served-model-name: "nvidia/MiniMax-M3-NVFP4"
tensor-parallel-size: 4
pipeline-parallel-size: 1
trust-remote-code: true
enable-prefix-caching: true
kv-cache-metrics: true
attention-config: '{"backend":"FLASHINFER","use_trtllm_attention":true,"indexer_kv_dtype":"fp8"}'
block-size: 128
gpu-memory-utilization: 0.9
max-model-len: 1048576
language-model-only: true
kv-cache-dtype: "fp8"
speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASHINFER"}'
kv-transfer-config: '{"kv_connector":"SimpleCPUOffloadConnector","kv_role":"kv_both","kv_connector_extra_config":{"cpu_bytes_to_use":777389080576,"cpu_bytes_to_use_per_rank":194347270144,"lazy_offload":true}}'
stream-interval: 20
compilation-config: '{"cudagraph_mode":"FULL_AND_PIECEWISE"}'
max-cudagraph-capture-size: 512
max-num-batched-tokens: 16384
max-num-seqs: 16
no-enable-flashinfer-autotune: true
reasoning-parser: "minimax_m3"
dyn-tool-call-parser: "minimax_m3"
dyn-reasoning-parser: "minimax_m3"

kv_events: true
sbatch_directives: {cpus-per-task: "144", mem: "0"}
srun_options: {container-remap-root: ""}

benchmark:
type: custom
command: bash /infmax-workspace/benchmarks/srt_agentic.sh
env:
INFMAX_CONTAINER_WORKSPACE: "/infmax-workspace"
RESULT_DIR: "/logs/agentic"
PORT: "8000"
IS_MULTINODE: "true"
AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: "true"
AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: "0"
AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: "14400"
AIPERF_EXTRA_INPUTS: "thinking:true"
AIPERF_REQUIRED_SERVER_METRIC_PREFIX: "vllm:"
AIPERF_DATASET_MMAP_CACHE_DIR: "/aiperf_mmap_cache"
HF_HUB_CACHE: "/hf_hub_cache"
WEKA_LOADER_OVERRIDE: "semianalysis_cc_traces_weka_062126"
109 changes: 109 additions & 0 deletions ...lurm-recipes/minimaxm3/vllm/gb200-fp4/agentx/agg-tp4-vllm-simple-mns8-nightly-native.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,109 @@
schema: 2
name: "minimax-m3-vllm-agg-gb200-tp4-vllm-simple-mns8-agentic-nightly-native"

model:
path: "minimax-m3-nvfp4"
container: "vllm/vllm-openai:nightly-9ea8f3ffc354901b740f0b31988900897b7221d7"
precision: "fp4"

identity:
model: {repo: "nvidia/MiniMax-M3-NVFP4"}
container: {image: "vllm/vllm-openai:nightly-9ea8f3ffc354901b740f0b31988900897b7221d7"}
frameworks: {dynamo: "1.5.0.dev20260908"}

dynamo: {install: true, source: {pypi: "1.5.0.dev20260908"}}
environment: {ETCD_LEASE_TTL: "7200"}
slurm: {time_limit: "12:00:00"}
health_check: {max_attempts: 2160, interval_seconds: 10}
resources: {gpu_type: "gb200", gpus_per_node: 4}
services:
- name: etcd
type: etcd
placement:
node: infra
- name: nats
type: nats
placement:
node: infra
options:
max_payload_mb: 32

frontend:
type: dynamo
enable_multiple_frontends: false
env: {DYN_TCP_REQUEST_TIMEOUT: "60"}
args:
trust-remote-code: true
router-mode: "kv"
router-kv-events: true
router-temperature: "0"
router-session-affinity-ttl-secs: 14400
kv-cache-block-size: 128

engine:
type: vllm
connector:
roles:
agg:
nodes: 1
workers: 1
gpus: 4
env:
VLLM_ENGINE_READY_TIMEOUT_S: "7200"
VLLM_EXECUTE_MODEL_TIMEOUT_SECONDS: "1800"
VLLM_FLOAT32_MATMUL_PRECISION: "high"
VLLM_FLASHINFER_ALLREDUCE_BACKEND: "trtllm"
VLLM_LOG_STATS_INTERVAL: "1"
VLLM_USE_SIMPLE_KV_OFFLOAD: "1"
NCCL_CUMEM_ENABLE: "1"
NCCL_MNNVL_ENABLE: "1"
NCCL_NVLS_ENABLE: "1"
UCX_MEMTYPE_CACHE: "n"
UCX_NET_DEVICES: "mlx5_0:1,mlx5_1:1,mlx5_2:1,mlx5_3:1"
UCX_TLS: "cuda_copy,cuda_ipc,rc"
NCCL_IB_HCA: "mlx5_0,mlx5_1,mlx5_2,mlx5_3"
args:
served-model-name: "nvidia/MiniMax-M3-NVFP4"
tensor-parallel-size: 4
pipeline-parallel-size: 1
trust-remote-code: true
enable-prefix-caching: true
kv-cache-metrics: true
attention-config: '{"backend":"FLASHINFER","use_trtllm_attention":true,"indexer_kv_dtype":"fp8"}'
block-size: 128
gpu-memory-utilization: 0.9
max-model-len: 1048576
language-model-only: true
kv-cache-dtype: "fp8"
speculative-config: '{"method":"eagle3","model":"Inferact/MiniMax-M3-EAGLE3-GQA","num_speculative_tokens":3,"attention_backend":"FLASHINFER"}'
kv-transfer-config: '{"kv_connector":"SimpleCPUOffloadConnector","kv_role":"kv_both","kv_connector_extra_config":{"cpu_bytes_to_use":777389080576,"cpu_bytes_to_use_per_rank":194347270144,"lazy_offload":true}}'
stream-interval: 20
compilation-config: '{"cudagraph_mode":"FULL_AND_PIECEWISE"}'
max-cudagraph-capture-size: 512
max-num-batched-tokens: 16384
max-num-seqs: 8
no-enable-flashinfer-autotune: true
reasoning-parser: "minimax_m3"
dyn-tool-call-parser: "minimax_m3"
dyn-reasoning-parser: "minimax_m3"

kv_events: true
sbatch_directives: {cpus-per-task: "144", mem: "0"}
srun_options: {container-remap-root: ""}

benchmark:
type: custom
command: bash /infmax-workspace/benchmarks/srt_agentic.sh
env:
INFMAX_CONTAINER_WORKSPACE: "/infmax-workspace"
RESULT_DIR: "/logs/agentic"
PORT: "8000"
IS_MULTINODE: "true"
AIPERF_HTTP_X_DYNAMO_SESSION_ID_FROM_CORRELATION_ID: "true"
AIPERF_USE_DYNAMO_CONV_AWARE_ROUTING: "0"
AIPERF_DYNAMO_SESSION_TIMEOUT_SECONDS: "14400"
AIPERF_EXTRA_INPUTS: "thinking:true"
AIPERF_REQUIRED_SERVER_METRIC_PREFIX: "vllm:"
AIPERF_DATASET_MMAP_CACHE_DIR: "/aiperf_mmap_cache"
HF_HUB_CACHE: "/hf_hub_cache"
WEKA_LOADER_OVERRIDE: "semianalysis_cc_traces_weka_062126"
Loading
Loading