config(dsv4-fp4 agentic): run offloading=none with expanded concurrency sweeps

Comment out the cpu (and prior none) entries for dsv4-fp4-b200/b300-vllm-agentic
and run offloading=none only, expanding the concurrency lists: dense low-to-mid
ramp on plain-TP entries and a cliff-spanning high-end ramp on the DEP entries
(B300 mirrors B200).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: cquil11

.github/configs/nvidia-master.yaml

@@ -9270,11 +9270,14 @@ dsv4-fp4-b200-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 (B200 25332045030: TP=8 1..32 + DEP=8 16..128 all 100%).
       # Re-add when investigating regressions in offload=none.
-      - { tp: 8, offloading: cpu,  conc-list: [16, 32, 64] }
-      - { tp: 8, ep: 8, dp-attn: true, offloading: cpu,  conc-list: [64, 128, 256] }
+      # - { tp: 8, offloading: cpu,  conc-list: [16, 32, 64] }
+      # - { tp: 8, ep: 8, dp-attn: true, offloading: cpu,  conc-list: [64, 128, 256] }
+      - { tp: 8, offloading: none,  conc-list: [1, 4, 8, 16, 32, 40, 48, 52, 64, 72] }
+      - { tp: 8, ep: 8, dp-attn: true, offloading: none,  conc-list: [52, 64, 72, 84, 100, 128, 196, 256, 512] }
 
 qwen3.5-fp8-b200-sglang-agentic:
   image: lmsysorg/sglang:nightly-dev-20260422-de962f32
@@ -9416,16 +9419,21 @@ 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: 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: 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: 8, ep: 8, dp-attn: true, offloading: none,  conc-list: [52, 64, 72, 84, 100, 128, 196, 256, 512] }
 
 gptoss-fp4-b200-vllm-agentic:
   image: vllm/vllm-openai:v0.22.0