[Bug] Llama4 Large Cuda Graph Memory #8538
Description
Checklist
- 1. I have searched related issues but cannot get the expected help.
- 2. The bug has not been fixed in the latest version.
- 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.
- 4. If the issue you raised is not a bug but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussions/new/choose Otherwise, it will be closed.
- 5. Please use English, otherwise it will be closed.
Describe the bug
When looking into #8468 , we observed high CUDA graph memory when launching Llama4-Maverick in SGLang. The total memory usage during cuda graph capture, as shown under the log is 6.19GB.
Version: 7c96971
See
[2025-07-29 18:14:43 TP0] Capture cuda graph end. Time elapsed: 19.50 s. mem usage=6.19 GB. avail mem=5.24 GB.
Full log at: (Added log info at 5bbc13b to show FA3 metadata memory)
Comparison with vLLM
- However, vLLM v0.9.2 only takes 1.40GiB
See:
(VllmWorker rank=0 pid=401) INFO 07-29 11:31:40 [gpu_model_runner.py:2326] Graph capturing finished in 52 secs, took 1.40 GiB
Full log at:
- By default, vLLM captures cuda graph with max batch size of 512, SGLang captures at 160.
Reproduction
sgl start command:
python3 -m sglang.launch_server --model /models/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 --port=8080 --tp-size=8 --mem-fraction-static=0.8 --context-length=524288 --tool-call-parser=pythonic --chat-template=examples/chat_template/tool_chat_template_llama4_pythonic.jinja --attention-backend=fa3 --enable-multimodal
vllm start command:
docker run -itd --gpus \"device=0,1,2,3,4,5,6,7\" \
--shm-size 10g \
--env HF_TOKEN=$<your-hf-token> \
--ulimit nofile=65535:65535 \
--network host \
--name vllm-prod-0.9.2-llama4-maverick \
vllm/vllm-openai:v0.9.2 \
--model=meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 \
--served-model-name=vllm-model \
--tensor-parallel-size=8 \
--max-model-len=524288 \
--limit-mm-per-prompt=image=10
Environment
Python: 3.12.3 (main, May 26 2025, 18:50:19) [GCC 13.3.0]
CUDA available: True
GPU 0,1,2,3,4,5,6,7: NVIDIA H100 80GB HBM3
GPU 0,1,2,3,4,5,6,7 Compute Capability: 9.0
CUDA_HOME: /usr/local/cuda
NVCC: Cuda compilation tools, release 12.6, V12.6.85
CUDA Driver Version: 560.35.05
PyTorch: 2.7.1+cu126
sglang: 0.4.9.post4
sgl_kernel: 0.2.7
flashinfer_python: 0.2.9rc2
triton: 3.3.1
transformers: 4.53.2
torchao: 0.9.0
numpy: 2.3.0
aiohttp: 3.12.13
fastapi: 0.115.12
hf_transfer: 0.1.9
huggingface_hub: 0.34.1
interegular: 0.3.3
modelscope: 1.27.0
orjson: 3.10.18
outlines: 0.1.11
packaging: 25.0
psutil: 7.0.0
pydantic: 2.11.7
python-multipart: 0.0.20
pyzmq: 27.0.0
uvicorn: 0.34.3
uvloop: 0.21.0
vllm: Module Not Found
xgrammar: 0.1.21
openai: 1.87.0
tiktoken: 0.9.0
anthropic: 0.54.0
litellm: 1.72.6
decord: 0.6.0
NVIDIA Topology:
GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 NIC8 NIC9 NIC10 NIC11 NIC12 NIC13 NIC14 NIC15 NIC16 NIC17 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 PXB PXB NODE NODE NODE NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS 0-55,112-167 0 N/A
GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 NODE NODE NODE PXB PXB NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS 0-55,112-167 0 N/A
GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 NODE NODE NODE NODE NODE PXB PXB NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS 0-55,112-167 0 N/A
GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 NODE NODE NODE NODE NODE NODE NODE PXB PXB SYS SYS SYS SYS SYS SYS SYS SYS SYS 0-55,112-167 0 N/A
GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 SYS SYS SYS SYS SYS SYS SYS SYS SYS PXB PXB NODE NODE NODE NODE NODE NODE NODE 56-111,168-223 1 N/A
GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE PXB PXB NODE NODE NODE NODE 56-111,168-223 1 N/A
GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE PXB PXB NODE NODE 56-111,168-223 1 N/A
GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE NODE NODE PXB PXB 56-111,168-223 1 N/A
NIC0 PXB NODE NODE NODE SYS SYS SYS SYS X PIX NODE NODE NODE NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC1 PXB NODE NODE NODE SYS SYS SYS SYS PIX X NODE NODE NODE NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC2 NODE NODE NODE NODE SYS SYS SYS SYS NODE NODE X NODE NODE NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC3 NODE PXB NODE NODE SYS SYS SYS SYS NODE NODE NODE X PIX NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC4 NODE PXB NODE NODE SYS SYS SYS SYS NODE NODE NODE PIX X NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC5 NODE NODE PXB NODE SYS SYS SYS SYS NODE NODE NODE NODE NODE X PIX NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC6 NODE NODE PXB NODE SYS SYS SYS SYS NODE NODE NODE NODE NODE PIX X NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC7 NODE NODE NODE PXB SYS SYS SYS SYS NODE NODE NODE NODE NODE NODE NODE X PIX SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC8 NODE NODE NODE PXB SYS SYS SYS SYS NODE NODE NODE NODE NODE NODE NODE PIX X SYS SYS SYS SYS SYS SYS SYS SYS SYS
NIC9 SYS SYS SYS SYS PXB NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS X PIX NODE NODE NODE NODE NODE NODE NODE
NIC10 SYS SYS SYS SYS PXB NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS PIX X NODE NODE NODE NODE NODE NODE NODE
NIC11 SYS SYS SYS SYS NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE X NODE NODE NODE NODE NODE NODE
NIC12 SYS SYS SYS SYS NODE PXB NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE X PIX NODE NODE NODE NODE
NIC13 SYS SYS SYS SYS NODE PXB NODE NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE PIX X NODE NODE NODE NODE
NIC14 SYS SYS SYS SYS NODE NODE PXB NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE X PIX NODE NODE
NIC15 SYS SYS SYS SYS NODE NODE PXB NODE SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE PIX X NODE NODE
NIC16 SYS SYS SYS SYS NODE NODE NODE PXB SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE NODE NODE X PIX
NIC17 SYS SYS SYS SYS NODE NODE NODE PXB SYS SYS SYS SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE NODE NODE PIX X
Legend:
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
NIC Legend:
NIC0: mlx5_0
NIC1: mlx5_1
NIC2: mlx5_2
NIC3: mlx5_3
NIC4: mlx5_4
NIC5: mlx5_5
NIC6: mlx5_6
NIC7: mlx5_7
NIC8: mlx5_8
NIC9: mlx5_9
NIC10: mlx5_10
NIC11: mlx5_11
NIC12: mlx5_12
NIC13: mlx5_13
NIC14: mlx5_14
NIC15: mlx5_15
NIC16: mlx5_16
NIC17: mlx5_17
ulimit soft: 65535