agentx: reshape kimik2.7 sweep to Pareto frontier + address review

- Frontier-only sweep: none [1,4,8] (low conc, fits GPU HBM) + lmcache [16,32]
  (above the KV cliff). Drops dominated/cliff points per review.
- Fix recipe header: 'Variant of kimik2.5_fp4_mi355x.sh' (was self-referential)
- perf-changelog: correct runner (cluster:mi355x-amds) and conc lists

Author: hyukjlee

benchmarks/single_node/agentic/kimik2.7_fp4_mi355x.sh

@@ -4,7 +4,7 @@ set -x
 
 # Agentic trace replay benchmark for Kimi-K2.7 FP4 on MI355X using vLLM.
 #
-# Variant of kimik2.7_fp4_mi355x.sh that supports TWO KV configs:
+# Variant of kimik2.5_fp4_mi355x.sh that supports TWO KV configs:
 #   KV_OFFLOADING=none                              -> GPU KV only
 #   KV_OFFLOADING=dram KV_OFFLOAD_BACKEND=lmcache   -> LMCache MP server + connector
 #

configs/amd-master.yaml

@@ -560,8 +560,8 @@ kimik2.7-fp4-mi355x-vllm-agentic:
     agentic-coding:
     - dram-utilization: 0.80
       search-space:
-      - { tp: 4, kv-offloading: none, conc-list: [1, 4, 8, 16, 32, 42] }
-      - { tp: 4, kv-offloading: dram, kv-offload-backend: lmcache, conc-list: [1, 4, 8, 16, 32, 42] }
+      - { tp: 4, kv-offloading: none, conc-list: [1, 4, 8] }
+      - { tp: 4, kv-offloading: dram, kv-offload-backend: lmcache, conc-list: [16, 32] }
 
 
 kimik2.5-fp4-mi355x-atom:

perf-changelog.yaml

@@ -5006,8 +5006,8 @@
 - config-keys:
     - kimik2.7-fp4-mi355x-vllm-agentic
   description:
-    - "Add Kimi-K2.7-Code-MXFP4 agentic-coding config on MI355X (cluster:gbt350docker->mi355x-amds) via vLLM v0.24.0, TP4"
-    - "Sweep KV-offloading none vs dram/LMCache at conc [1,4,8,16,32,42]; LMCache MP server + LMCacheMPConnector, dram-utilization 0.80"
+    - "Add Kimi-K2.7-Code-MXFP4 agentic-coding config on MI355X (runner: cluster:mi355x-amds) via vLLM v0.24.0, TP4"
+    - "Sweep KV-offloading none (conc [1,4,8]) vs dram/LMCache (conc [16,32]); LMCache MP server + LMCacheMPConnector, dram-utilization 0.80"
     - "Recipe benchmarks/single_node/agentic/kimik2.7_fp4_mi355x.sh clones LMCache to container-local /opt/lmcache-src (pinned) so CI checkout never trips over root-owned build artifacts"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2126