Skip to content

[NVIDIA] Update H200/B200 SGLang image to v0.5.5-cu129-amd64 and fix deprecated flags #204

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

Merged
functionstackx merged 5 commits into from
Nov 10, 2025
Merged

[NVIDIA] Update H200/B200 SGLang image to v0.5.5-cu129-amd64 and fix deprecated flags #204

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
623a624
Update h200/b200 sglang image tags to v0.5.5-cu129-amd64
Copilot Nov 9, 2025
d6f38f2
Fix deprecated SGLang flags: replace --enable-ep-moe with --ep-size 8…
Copilot Nov 9, 2025
d1d2c82
Use $EP_SIZE variable instead of hardcoded 8 and add ep: 8 to nvidia-…
Copilot Nov 9, 2025
a21de78
Add ep: 4 for tp=4 entries in dsr1-fp4-b200-sglang config
Copilot Nov 9, 2025
28534c7
Pass EP_SIZE environment variable to Docker containers in B200 runner…
Copilot Nov 9, 2025
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
24 changes: 12 additions & 12 deletions .github/configs/nvidia-master.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
dsr1-fp4-b200-sglang:
image: lmsysorg/sglang:v0.5.3rc1-cu129-b200
image: lmsysorg/sglang:v0.5.5-cu129-amd64
model: nvidia/DeepSeek-R1-0528-FP4-V2
model-prefix: dsr1
runner: b200
Expand All @@ -9,18 +9,18 @@ dsr1-fp4-b200-sglang:
- isl: 1024
osl: 1024
search-space:
- { tp: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, conc-start: 4, conc-end: 128 }
- { tp: 4, ep: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 128 }
- isl: 1024
osl: 8192
search-space:
- { tp: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, conc-start: 4, conc-end: 128 }
- { tp: 4, ep: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 128 }
- isl: 8192
osl: 1024
search-space:
- { tp: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, conc-start: 4, conc-end: 16 }
- { tp: 4, ep: 4, conc-start: 4, conc-end: 128 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 16 }

dsr1-fp4-b200-trt:
image: nvcr.io#nvidia/tensorrt-llm/release:1.1.0rc2.post2
Expand Down Expand Up @@ -73,7 +73,7 @@ dsr1-fp4-b200-trt:
- { tp: 8, ep: 8, dp-attn: true, conc-start: 64, conc-end: 256 }

dsr1-fp8-b200-sglang:
image: lmsysorg/sglang:v0.5.3rc1-cu129-b200
image: lmsysorg/sglang:v0.5.5-cu129-amd64
model: deepseek-ai/DeepSeek-R1-0528
model-prefix: dsr1
runner: b200
Expand All @@ -83,15 +83,15 @@ dsr1-fp8-b200-sglang:
- isl: 1024
osl: 1024
search-space:
- { tp: 8, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 64 }
- isl: 1024
osl: 8192
search-space:
- { tp: 8, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 64 }
- isl: 8192
osl: 1024
search-space:
- { tp: 8, conc-start: 4, conc-end: 64 }
- { tp: 8, ep: 8, conc-start: 4, conc-end: 64 }

dsr1-fp8-b200-trt:
image: nvcr.io#nvidia/tensorrt-llm/release:1.1.0rc2.post2
Expand Down Expand Up @@ -120,7 +120,7 @@ dsr1-fp8-b200-trt:
- { tp: 8, ep: 8, conc-start: 4, conc-end: 64 }

dsr1-fp8-h200-sglang:
image: lmsysorg/sglang:v0.5.2rc2-cu126
image: lmsysorg/sglang:v0.5.5-cu129-amd64
model: deepseek-ai/DeepSeek-R1-0528
model-prefix: dsr1
runner: h200
Expand Down
4 changes: 2 additions & 2 deletions benchmarks/dsr1_fp4_b200_docker.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,6 @@ PYTHONNOUSERSITE=1 python3 -m sglang.launch_server --model-path $MODEL --host 0.
--tensor-parallel-size=$TP --data-parallel-size=1 \
--cuda-graph-max-bs 256 --max-running-requests 256 --mem-fraction-static 0.85 --kv-cache-dtype fp8_e4m3 \
--chunked-prefill-size 16384 \
--enable-ep-moe --quantization modelopt_fp4 --enable-flashinfer-allreduce-fusion --scheduler-recv-interval $SCHEDULER_RECV_INTERVAL \
--enable-symm-mem --disable-radix-cache --attention-backend trtllm_mla --enable-flashinfer-trtllm-moe --stream-interval 10
--ep-size $EP_SIZE --quantization modelopt_fp4 --enable-flashinfer-allreduce-fusion --scheduler-recv-interval $SCHEDULER_RECV_INTERVAL \
--enable-symm-mem --disable-radix-cache --attention-backend trtllm_mla --moe-runner-backend flashinfer_trtllm --stream-interval 10

2 changes: 1 addition & 1 deletion benchmarks/dsr1_fp8_b200_docker.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,4 +34,4 @@ PYTHONNOUSERSITE=1 python3 -m sglang.launch_server --model-path=$MODEL --host=0.
--cuda-graph-max-bs 128 --max-running-requests 128 \
--mem-fraction-static 0.82 --kv-cache-dtype fp8_e4m3 --chunked-prefill-size 32768 --max-prefill-tokens 32768 \
--enable-flashinfer-allreduce-fusion --scheduler-recv-interval $SCHEDULER_RECV_INTERVAL --disable-radix-cache \
--attention-backend trtllm_mla --stream-interval 30 --enable-flashinfer-trtllm-moe --quantization fp8
--attention-backend trtllm_mla --stream-interval 30 --ep-size $EP_SIZE --moe-runner-backend flashinfer_trtllm --quantization fp8
2 changes: 1 addition & 1 deletion runners/launch_b200-nvd.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ docker run --rm -d --init --network host --name $server_name \
--runtime nvidia --gpus all --ipc host --privileged --shm-size=16g --ulimit memlock=-1 --ulimit stack=67108864 \
-v $HF_HUB_CACHE_MOUNT:$HF_HUB_CACHE \
-v $GITHUB_WORKSPACE:/workspace/ -w /workspace/ \
-e HF_TOKEN -e HF_HUB_CACHE -e MODEL -e TP -e CONC -e MAX_MODEL_LEN -e ISL -e OSL -e PORT=$PORT \
-e HF_TOKEN -e HF_HUB_CACHE -e MODEL -e TP -e CONC -e MAX_MODEL_LEN -e ISL -e OSL -e PORT=$PORT -e EP_SIZE \
-e NCCL_GRAPH_REGISTER=0 \
-e TORCH_CUDA_ARCH_LIST="10.0" -e CUDA_DEVICE_ORDER=PCI_BUS_ID -e CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" \
--entrypoint=/bin/bash \
Expand Down
2 changes: 1 addition & 1 deletion runners/launch_b200-tg.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@ docker run --rm -d --network host --name $server_name \
--runtime nvidia --gpus all --ipc host --privileged --shm-size=16g --ulimit memlock=-1 --ulimit stack=67108864 \
-v $HF_HUB_CACHE_MOUNT:$HF_HUB_CACHE \
-v $GITHUB_WORKSPACE:/workspace/ -w /workspace/ \
-e HF_TOKEN -e HF_HUB_CACHE -e MODEL -e TP -e CONC -e MAX_MODEL_LEN -e ISL -e OSL -e PORT=$PORT \
-e HF_TOKEN -e HF_HUB_CACHE -e MODEL -e TP -e CONC -e MAX_MODEL_LEN -e ISL -e OSL -e PORT=$PORT -e EP_SIZE \
-e TORCH_CUDA_ARCH_LIST="10.0" -e CUDA_DEVICE_ORDER=PCI_BUS_ID -e CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" \
--entrypoint=/bin/bash \
$(echo "$IMAGE" | sed 's/#/\//') \
Expand Down
Loading