(testing) add offload off scneario to dsv4 b300

Author: cquil11

.github/configs/nvidia-master.yaml

@@ -9259,7 +9259,7 @@ glm5-fp8-gb300-dynamo-sglang:
           dp-attn: false
 
 dsv4-fp4-b200-vllm-agentic:
-  image: cquil/vllm-openai:v0.22.0-6c529f3001ab8bf44b1657e779dc54b622397045
+  image: vllm/vllm-openai:nightly-d0975a4b50140a9d953f00955a1cbb2a4945edef
   model: deepseek-ai/DeepSeek-V4-Pro
   model-prefix: dsv4
   runner: b200-dgxc
@@ -9408,7 +9408,8 @@ dsv4-fp8-h200-vllm-agentic:
 dsv4-fp4-b300-vllm-agentic:
   # image: vllm/vllm-openai:v0.22.0
   # includes https://github.com/vllm-project/vllm/pull/43447 up to 6c529f3001ab8bf44b1657e779dc54b622397045
-  image: cquil/vllm-openai:v0.22.0-6c529f3001ab8bf44b1657e779dc54b622397045
+  # image: cquil/vllm-openai:v0.22.0-6c529f3001ab8bf44b1657e779dc54b622397045'
+  image: vllm/vllm-openai:nightly-d0975a4b50140a9d953f00955a1cbb2a4945edef
   model: deepseek-ai/DeepSeek-V4-Pro
   model-prefix: dsv4
   runner: b300
@@ -9419,20 +9420,10 @@ dsv4-fp4-b300-vllm-agentic:
     agentic-coding:
     - duration: 1800
       search-space:
-      # TEMPORARILY COMMENTED OUT — running offloading=none only this iteration.
-      # cpu offload only this iteration — none entries already validated in
-      # earlier runs. Re-add when investigating regressions in offload=none.
-      # - { tp: 4, offloading: cpu,  conc-list: [16, 32, 64] }
-      # - { tp: 8, offloading: cpu,  conc-list: [16, 32, 64] }
-      # - { tp: 4, ep: 4, dp-attn: true, offloading: cpu,  conc-list: [64, 128, 256] }
-      # - { tp: 8, ep: 8, dp-attn: true, offloading: cpu,  conc-list: [128, 256, 512] }
-      # - { tp: 4, offloading: none,  conc-list: [16, 32, 64] }
-      # - { tp: 8, offloading: none,  conc-list: [16, 32, 64] }
-      # - { tp: 4, ep: 4, dp-attn: true, offloading: none,  conc-list: [64, 128, 256] }
-      # - { tp: 8, ep: 8, dp-attn: true, offloading: none,  conc-list: [128, 256, 512] }
-      - { tp: 4, offloading: none,  conc-list: [1, 4, 8, 16, 32, 40, 48, 52, 64, 72] }
+      - { tp: 4, offloading: none,  conc-list: [1, 4, 8, 16, 32] }
       - { tp: 8, offloading: none,  conc-list: [1, 4, 8, 16, 32, 40, 48, 52, 64, 72] }
-      - { tp: 4, ep: 4, dp-attn: true, offloading: none,  conc-list: [52, 64, 72, 84, 100, 128, 196, 256, 512] }
+      - { tp: 4, ep: 4, dp-attn: true, offloading: none,  conc-list: [8, 16, 32, 64, 128] }
+      - { tp: 4, ep: 4, dp-attn: true, offloading: cpu,   conc-list: [32, 48, 64, 96, 128, 192, 256] }
       - { tp: 8, ep: 8, dp-attn: true, offloading: none,  conc-list: [52, 64, 72, 84, 100, 128, 196, 256, 512] }
 
 gptoss-fp4-b200-vllm-agentic: