Use $EP_SIZE variable instead of hardcoded 8 and add ep: 8 to nvidia-master.yaml for B200 SGLang configs

Co-authored-by: functionstackx <47992694+functionstackx@users.noreply.github.com>

Author: Copilot

.github/configs/nvidia-master.yaml

@@ -10,17 +10,17 @@ dsr1-fp4-b200-sglang:
     osl: 1024
     search-space:
     - { tp: 4, conc-start: 4, conc-end: 128 }
-    - { tp: 8, 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: 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: 8, ep: 8, conc-start: 4, conc-end: 16 }
 
 dsr1-fp4-b200-trt:
   image: nvcr.io#nvidia/tensorrt-llm/release:1.1.0rc2.post2
@@ -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

benchmarks/dsr1_fp4_b200_docker.sh

@@ -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 \
---ep-size 8 --quantization modelopt_fp4 --enable-flashinfer-allreduce-fusion --scheduler-recv-interval $SCHEDULER_RECV_INTERVAL \
+--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
 

benchmarks/dsr1_fp8_b200_docker.sh

@@ -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 --moe-runner-backend flashinfer_trtllm --quantization fp8
+--attention-backend trtllm_mla --stream-interval 30 --ep-size $EP_SIZE --moe-runner-backend flashinfer_trtllm --quantization fp8