Add dsr1 and gpt-oss test cases

Signed-off-by: Chenfei Zhang <[email protected]>

Author: chenfeiz0326

jenkins/L0_Test.groovy

@@ -2698,6 +2698,7 @@ def launchTestJobs(pipeline, testFilter)
         "DGX_B300-4_GPUs-PyTorch-Post-Merge-1": ["b300-x4", "l0_dgx_b300", 1, 1, 4],
         // Perf sanity post merge test
         "DGX_B200-4_GPUs-PyTorch-Perf-Sanity-Post-Merge-1": ["b200-x4", "perf_sanity_l0_dgx_b200", 1, 1, 4],
+        "DGX_B200-8_GPUs-PyTorch-Perf-Sanity-Post-Merge-1": ["b200-x8", "perf_sanity_l0_dgx_b200", 1, 1, 8],
         "DGX_B300-4_GPUs-PyTorch-Perf-Sanity-Post-Merge-1": ["b300-x4", "perf_sanity_l0_dgx_b300", 1, 1, 4],
     ]
     fullSet += x86SlurmTestConfigs.keySet()
@@ -2730,6 +2731,7 @@ def launchTestJobs(pipeline, testFilter)
         // "GB300-PyTorch-1": ["gb300-single", "l0_gb300", 1, 1],
         "GB200-4_GPUs-PyTorch-1": ["gb200-trtllm", "l0_gb200_multi_gpus", 1, 1, 4],
         "GB200-4_GPUs-PyTorch-Post-Merge-1": ["gb200-trtllm", "l0_gb200_multi_gpus", 1, 1, 4],
+        "GB200-4_GPUs-PyTorch-Perf-Sanity-Post-Merge-1": ["gb200-trtllm", "perf_sanity_l0_gb200_multi_gpus", 1, 1, 4],
         // "GB300-4_GPUs-PyTorch-Post-Merge-1": ["gb300-trtllm", "l0_gb300_multi_gpus", 1, 1, 4],
     ]
     fullSet += SBSASlurmTestConfigs.keySet()

tests/integration/defs/perf/test_perf.py

@@ -687,7 +687,7 @@ def generate_extra_llm_api_config(self) -> str:
                 config_lines.append(f"  max_draft_len: {self.max_draft_len}")
             if self.speculative_model_dir:
                 config_lines.append(
-                    f"  speculative_model_dir: {speculative_model_dir}")
+                    f"  speculative_model_dir: {self.speculative_model_dir}")
 
         return "\n".join(config_lines)
 

tests/integration/test_lists/test-db/perf_sanity_l0_dgx_b200.yml

@@ -15,7 +15,7 @@ perf_sanity_l0_dgx_b200:
       backend: pytorch
       orchestrator: mpi
   tests:
-  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200]
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200-r1_fp4_dep4,r1_fp4_tep4,r1_fp4_v2_dep4_mtp1,r1_fp4_v2_tep4_mtp3,gpt_oss_dep4,gpt_oss_tep4]
 
 - condition:
     ranges:
@@ -32,4 +32,38 @@ perf_sanity_l0_dgx_b200:
       backend: pytorch
       orchestrator: mpi
   tests:
-  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200]
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200-r1_fp4_dep4,r1_fp4_tep4,r1_fp4_v2_dep4_mtp1,r1_fp4_v2_tep4_mtp3,gpt_oss_dep4,gpt_oss_tep4]
+
+- condition:
+    ranges:
+      system_gpu_count:
+        gte: 8
+        lte: 8
+    wildcards:
+      gpu:
+      - '*b200*'
+      linux_distribution_name: ubuntu*
+      cpu: x86_64
+    terms:
+      stage: pre_merge
+      backend: pytorch
+      orchestrator: mpi
+  tests:
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200-r1_fp8_dep4_mtp1,r1_fp8_tep4_mtp3]
+
+- condition:
+    ranges:
+      system_gpu_count:
+        gte: 8
+        lte: 8
+    wildcards:
+      gpu:
+      - '*b200*'
+      linux_distribution_name: ubuntu*
+      cpu: x86_64
+    terms:
+      stage: post_merge
+      backend: pytorch
+      orchestrator: mpi
+  tests:
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_dgx_b200-r1_fp8_dep4_mtp1,r1_fp8_tep4_mtp3]

tests/integration/test_lists/test-db/perf_sanity_l0_gb200_multi_gpus.yml

@@ -0,0 +1,35 @@
+version: 0.0.1
+perf_sanity_l0_gb200_multi_gpus:
+- condition:
+    ranges:
+      system_gpu_count:
+        gte: 4
+        lte: 4
+    wildcards:
+      gpu:
+      - '*gb200*'
+      linux_distribution_name: ubuntu*
+      cpu: aarch64
+    terms:
+      stage: pre_merge
+      backend: pytorch
+      orchestrator: mpi
+  tests:
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_gb200_multi_gpus]
+
+- condition:
+    ranges:
+      system_gpu_count:
+        gte: 4
+        lte: 4
+    wildcards:
+      gpu:
+      - '*gb200*'
+      linux_distribution_name: ubuntu*
+      cpu: aarch64
+    terms:
+      stage: post_merge
+      backend: pytorch
+      orchestrator: mpi
+  tests:
+  - perf/test_perf.py::test_perf[perf_sanity_upload-l0_gb200_multi_gpus]

tests/scripts/perf-sanity/l0_dgx_gb200.yaml renamed to tests/scripts/perf-sanity/l0_gb200_multi_gpus.yaml

@@ -32,7 +32,7 @@ server_configs:
   - name: "r1_fp4_v2_tep4_mtp3"
     model_name: "deepseek_r1_0528_fp4_v2"
     gpus: 4
-    tp: 4W
+    tp: 4
     ep: 1
     pp: 1
     attention_backend: "TRTLLM"