waive llama_v3_405b in perf test

Signed-off-by: ruodil <[email protected]>

Author: ruodil

tests/integration/test_lists/qa/trt_llm_release_perf_cluster_test.yml

@@ -21,7 +21,6 @@ trt_llm_release_perf_cluster_test:
   - perf/test_perf.py::test_perf[llama_v3.3_70b_instruct_fp8-bench-float8-maxbs:256-input_output_len:128,128-gpus:2]
   - perf/test_perf.py::test_perf[llama_v3.3_70b_instruct_fp8-bench-float8-maxbs:256-input_output_len:512,32-gpus:2]
   - perf/test_perf.py::test_perf[llama_v2_13b-bench-float16-input_output_len:128,128-loras:8-gpus:2]
-  - perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:128,128-gpus:2]
   - perf/test_perf.py::test_perf[mixtral_8x7b_v0.1-bench-float16-input_output_len:128,128-quant:fp8-gpus:2]
 
 # Tests for systems with 4+ GPUs
@@ -34,16 +33,17 @@ trt_llm_release_perf_cluster_test:
   - perf/test_perf.py::test_perf[qwen_14b_chat-bench-float16-input_output_len:128,128-gpus:4]
   - perf/test_perf.py::test_perf[qwen_14b_chat-bench-float16-input_output_len:512,32-gpus:4]
   - perf/test_perf.py::test_perf[starcoder_15b-bench-float16-input_output_len:512,200-gpus:4]
-  - perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:128,128-gpus:4]
+  #- perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:128,128-gpus:4]
 
 # Tests for systems with 8+ GPUs
 - condition:
     ranges:
       system_gpu_count:
         gte: 8
   tests:
-  - perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:128,128-gpus:8]
-  - perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:512,32-gpus:8]
+  #- perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:128,128-gpus:8]
+  #- perf/test_perf.py::test_perf[llama_v3.1_405b_instruct_fp4-bench-pytorch-float4-input_output_len:512,32-gpus:8]
+  - perf/test_perf.py::test_perf[llama_v3.3_nemotron_super_49b-bench-bfloat16-input_output_len:500,2000-quant:fp8-con:250-gpus:8]
   - perf/test_perf.py::test_perf[mixtral_8x22b_v0.1-bench-float16-input_output_len:512,512-quant:fp8-tp:8]
   - perf/test_perf.py::test_perf[gpt_20b-bench-float16-maxbs:8-input_output_len:128,128-reqs:80-gpus:8]
   - perf/test_perf.py::test_perf[gpt_20b-bench-float16-maxbs:8-input_output_len:512,32-reqs:80-gpus:8]

tests/integration/test_lists/qa/trt_llm_release_perf_test.yml

@@ -90,7 +90,6 @@ trt_llm_release_perf_test:
   - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-pytorch-bfloat16-maxbs:512-input_output_len:1000,1000-reqs:8-con:1]
   - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-pytorch-bfloat16-maxbs:512-maxnt:20000-input_output_len:20000,2000-reqs:8-con:1]
   - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-pytorch-bfloat16-maxbs:512-input_output_len:500,2000-reqs:500-con:250]
-  - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-pytorch-bfloat16-maxbs:512-input_output_len:1000,1000-reqs:500-con:250]
   #long time llama_nemotron cases
   - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-bfloat16-maxbs:64-input_output_len:5000,500-con:1] # timeout for l20, l40s, a100
   - perf/test_perf.py::test_perf[llama_v3.1_nemotron_nano_8b-bench-bfloat16-maxbs:64-input_output_len:5000,500-quant:fp8-con:1] #timeout for l20, l40s, failed for a100
@@ -304,6 +303,7 @@ trt_llm_release_perf_test:
   # E2E trtllm-bench
   - perf/test_perf.py::test_perf[llama_v3.1_70b_instruct-cppmanager-exe-plugin_ifb-float16-input_output_len:200,2000-reqs:64-gpus:8]
   - perf/test_perf.py::test_perf[llama_v3.1_70b_instruct-bench-bfloat16-input_output_len:200,2000-reqs:64-con:200-gpus:8]
+  - perf/test_perf.py::test_perf[llama_v3.1_70b_instruct-bench-pytorch-bfloat16-input_output_len:200,2000-reqs:64-con:200-gpus:8]
   - perf/test_perf.py::test_perf[llama_v3.1_70b_instruct-bench-bfloat16-input_output_len:200,2000-reqs:64-con:1-gpus:8] # timeout for h20, move to l2 test
   - perf/test_perf.py::test_perf[llama_v3.1_70b_instruct-bench-bfloat16-input_output_len:2000,200-reqs:64-gpus:8]
   - perf/test_perf.py::test_perf[llama_v3.3_70b_instruct_fp8-bench-pytorch-float8-input_output_len:128,128-gpus:8]