Fix test case as chunked attention not Supported on sm_120

Author: yufeiwu-nv

tests/integration/test_lists/qa/llm_perf_core.yml

@@ -274,15 +274,14 @@ llm_perf_core:
 
 - condition:
     ranges:
+      compute_capability:
+        gte: 9.0
+        lt: 12.0
       system_gpu_count:
         gte: 8
       gpu_memory:
         gt: 80000
-    wildcards:
-      gpu:
-      - '*h100*'
-      - '*h200*'
-      - '*h20*'
+
   tests:
   # E2E trtllm-bench
   #mixtral_8x7b_v0.1_instruct
@@ -309,7 +308,7 @@ llm_perf_core:
   - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-maxbs:1024-maxnt:4096-kv_frac:0.85-input_output_len:1000,1000-reqs:3000-ep:8-tp:8-gpus:8]
   - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-input_output_len:128,128-ep:8-tp:8-gpus:8]
   - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-input_output_len:512,32-ep:8-tp:8-gpus:8]
-  #rcca case
+  # chunked attention case
   - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-maxbs:1024-maxnt:20000-kv_frac:0.6-input_output_len:20000,2000-reqs:1000-ep:8-tp:8-gpus:8]
 
   #llama_v4_scout_17b_16e_instruct_fp8

tests/integration/test_lists/qa/llm_perf_sanity.yml

@@ -168,11 +168,23 @@ llm_perf_sanity:
   # for chunked prefill cases
   - perf/test_perf.py::test_perf[deepseek_v3_lite_fp8-bench-pytorch-float8-maxbs:512-maxnt:2048-kv_frac:0.85-input_output_len:3000,500-reqs:200]
   - perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp8-bench-pytorch-float8-input_output_len:128,128-tp:8-gpus:8]
-  - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-input_output_len:512,32-ep:8-tp:8-gpus:8]
-  - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-maxbs:1024-maxnt:20000-kv_frac:0.6-input_output_len:20000,2000-reqs:1000-ep:8-tp:8-gpus:8]
+  - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-input_output_len:512,32-ep:8-tp:8-gpus:8] TIMEOUT(100)
   - perf/test_perf.py::test_perf[llama_v4_scout_17b_16e_instruct_fp8-bench-pytorch-float8-maxbs:1024-maxnt:20000-kv_frac:0.85-input_output_len:20000,2000-reqs:1000-ep:8-tp:8-gpus:8] TIMEOUT(100)
   - perf/test_perf.py::test_perf[qwen3_235b_a22b_fp8-bench-pytorch-float8-input_output_len:1000,2000-con:256-ep:8-gpus:8] TIMEOUT(60)
   - perf/test_perf.py::test_perf[deepseek_v3_lite_fp8-disagg_server-ctx_dp:4-gen_tp:4]
   - perf/test_perf.py::test_perf[llama_v3.1_8b-disagg_server-ctx_dp:4-gen_tp:4]
     # gpt_oss_20b_fp4
   - perf/test_perf.py::test_perf[gpt_oss_20b_fp4-bench-pytorch-float4-input_output_len:512,512]
+
+# gpu_arch > Hopper, exculde GB20X, RTX 6000 for not supported
+- condition:
+    ranges:
+      system_gpu_count:
+        gte: 8
+      compute_capability:
+        gte: 9.0
+        lt: 12.0
+
+  tests:
+  # chunked attention case
+  - perf/test_perf.py::test_perf[llama_v4_maverick_17b_128e_instruct_fp8-bench-pytorch-float8-maxbs:1024-maxnt:20000-kv_frac:0.6-input_output_len:20000,2000-reqs:1000-ep:8-tp:8-gpus:8]