[None][test] Consolidate test coverage for helix

tests/integration/defs/accuracy/test_disaggregated_serving.py

@@ -1212,30 +1212,19 @@ def test_auto_dtype(self, overlap_scheduler, mtp_nextn):
 
     @skip_pre_blackwell
     @pytest.mark.skip_less_device(8)
-    @pytest.mark.parametrize(
-        "gen_pp,gen_tp,gen_cp,enable_attention_dp", [
-            (1, 1, 4, False),
-            (1, 2, 2, False),
-            (1, 2, 2, True),
-            (2, 1, 2, False),
-        ],
-        ids=["pp1tp1cp4", "pp1tp2cp2", "pp1dp2cp2", "pp2tp1cp2"])
+    @pytest.mark.parametrize("gen_pp,gen_tp,gen_cp,enable_attention_dp", [
+        (1, 2, 2, False),
+        (1, 2, 2, True),
+    ],
+                             ids=["pp1tp2cp2", "pp1dp2cp2"])
     @pytest.mark.parametrize("cuda_graph_config", [
-        None,
-        {
-            "enable_padding": False,
-            "batch_sizes": [1, 2, 4, 8, 16, 32, 64]
-        },
         {
             "enable_padding": True,
             "batch_sizes": [1, 2, 4, 8, 16, 32, 64]
         },
     ],
-                             ids=[
-                                 "cudagraph:none", "cudagraph:without_padding",
-                                 "cudagraph:with_padding"
-                             ])
-    @pytest.mark.parametrize("comms_medium", ["fifo_v1", "fifo_v2", "nccl"])
+                             ids=["cudagraph:with_padding"])
+    @pytest.mark.parametrize("comms_medium", ["fifo_v2"])
     def test_auto_dtype_with_helix(self, comms_medium, cuda_graph_config,
                                    gen_pp, gen_tp, gen_cp, enable_attention_dp):
         # Parse comms_medium to get use_nccl_for_alltoall and fifo_version.
@@ -1763,102 +1752,6 @@ def test_auto_dtype(self, overlap_scheduler):
                               model_name=self.MODEL_NAME,
                               test_sets=["MMLU", "GSM8K"])
 
-    @skip_pre_blackwell
-    @pytest.mark.skip_less_device(8)
-    @pytest.mark.parametrize(
-        "gen_pp,gen_tp,gen_cp,enable_attention_dp", [
-            (1, 1, 4, False),
-            (1, 2, 2, False),
-            (1, 2, 2, True),
-            (2, 1, 2, False),
-        ],
-        ids=["pp1tp1cp4", "pp1tp2cp2", "pp1dp2cp2", "pp2tp1cp2"])
-    @pytest.mark.parametrize("cuda_graph_config", [
-        None,
-        {
-            "enable_padding": True,
-            "batch_sizes": [1, 2, 4, 8, 16, 32, 64],
-        },
-    ],
-                             ids=[
-                                 "cudagraph:none",
-                                 "cudagraph:with_padding",
-                             ])
-    @pytest.mark.parametrize("comms_medium", ["fifo", "nccl"])
-    def test_auto_dtype_with_helix(self, comms_medium, cuda_graph_config,
-                                   gen_pp, gen_tp, gen_cp, enable_attention_dp):
-        use_nccl_for_alltoall = comms_medium == "nccl"
-        fifo_version = 2
-        gen_ep = gen_tp * gen_cp
-        kv_cache_config = {
-            "free_gpu_memory_fraction": 0.5,
-            "enable_block_reuse": False,
-            "enable_partial_reuse": False,
-            "tokens_per_block": 32,
-            "dtype": "fp8",
-        }
-        ctx_server_config = {
-            "pipeline_parallel_size": 1,
-            "tensor_parallel_size": 4,
-            "context_parallel_size": 1,
-            "disable_overlap_scheduler": True,
-            "kv_cache_config": kv_cache_config,
-            "enable_chunked_prefill": False,
-            "cuda_graph_config": None,
-            "cache_transceiver_config": {
-                "backend": "UCX",
-                "max_tokens_in_buffer": 8192,
-            },
-            "moe_config": {
-                "backend": "TRTLLM",
-                "max_num_tokens": 16384,
-            },
-        }
-        gen_server_config = {
-            "tensor_parallel_size": gen_tp,
-            "pipeline_parallel_size": gen_pp,
-            "context_parallel_size": gen_cp,
-            "moe_expert_parallel_size": gen_ep,
-            "cp_config": {
-                "cp_type": "HELIX",
-                "tokens_per_block": 32,
-                "use_nccl_for_alltoall": use_nccl_for_alltoall,
-                "fifo_version": fifo_version,
-            },
-            "disable_overlap_scheduler": True,
-            "kv_cache_config": kv_cache_config,
-            "enable_chunked_prefill": False,
-            "cuda_graph_config": cuda_graph_config,
-            "cache_transceiver_config": {
-                "backend": "UCX",
-                "max_tokens_in_buffer": 8192,
-            },
-            "moe_config": {
-                "backend": "TRTLLM",
-                "max_num_tokens": 16384,
-            },
-            "enable_attention_dp": enable_attention_dp,
-        }
-        disaggregated_server_config = {
-            "hostname": "localhost",
-            "port": 8000,
-            "backend": "pytorch",
-            "context_servers": {
-                "num_instances": 1,
-                "urls": ["localhost:8001"]
-            },
-            "generation_servers": {
-                "num_instances": 1,
-                "urls": ["localhost:8002"]
-            }
-        }
-        with launch_disaggregated_llm(disaggregated_server_config,
-                                      ctx_server_config,
-                                      gen_server_config,
-                                      self.MODEL_PATH,
-                                      max_workers=128) as llm:
-            run_accuracy_test(llm, self.MODEL_NAME, ["GSM8K"])
-
 
 @pytest.mark.timeout(DEFAULT_TEST_TIMEOUT)
 @skip_pre_hopper
@@ -1988,30 +1881,19 @@ def test_chunked_prefill(self):
 
     @skip_pre_blackwell
     @pytest.mark.skip_less_device(8)
-    @pytest.mark.parametrize(
-        "gen_pp,gen_tp,gen_cp,enable_attention_dp", [
-            (1, 1, 4, False),
-            (1, 2, 2, False),
-            (1, 2, 2, True),
-            (2, 1, 2, False),
-        ],
-        ids=["pp1tp1cp4", "pp1tp2cp2", "pp1dp2cp2", "pp2tp1cp2"])
+    @pytest.mark.parametrize("gen_pp,gen_tp,gen_cp,enable_attention_dp", [
+        (1, 2, 2, False),
+        (1, 2, 2, True),
+    ],
+                             ids=["pp1tp2cp2", "pp1dp2cp2"])
     @pytest.mark.parametrize("cuda_graph_config", [
-        None,
-        {
-            "enable_padding": False,
-            "batch_sizes": [1, 2, 4, 8, 16, 32, 64]
-        },
         {
             "enable_padding": True,
             "batch_sizes": [1, 2, 4, 8, 16, 32, 64]
         },
     ],
-                             ids=[
-                                 "cudagraph:none", "cudagraph:without_padding",
-                                 "cudagraph:with_padding"
-                             ])
-    @pytest.mark.parametrize("comms_medium", ["fifo_v1", "fifo_v2", "nccl"])
+                             ids=["cudagraph:with_padding"])
+    @pytest.mark.parametrize("comms_medium", ["fifo_v2"])
     def test_auto_dtype_with_helix(self, comms_medium, cuda_graph_config,
                                    gen_pp, gen_tp, gen_cp, enable_attention_dp):
         # Parse comms_medium to get use_nccl_for_alltoall and fifo_version.

tests/integration/defs/disaggregated/test_disaggregated.py

@@ -2655,8 +2655,7 @@ def test_llama4_long_context_kv_cache_overflow(disaggregated_test_root,
 @skip_pre_blackwell
 @pytest.mark.timeout(2400)
 @pytest.mark.skip_less_device(4)
-@pytest.mark.parametrize("prompt_file", ["prompts.json", "long_prompts.json"],
-                         ids=["short_prompt", "long_prompt"])
+@pytest.mark.parametrize("prompt_file", ["prompts.json"], ids=["short_prompt"])
 @pytest.mark.parametrize("deepseek_v3_model_root", ['DeepSeek-V3-Lite-bf16'],
                          indirect=True)
 def test_disaggregated_deepseek_v3_lite_bf16_tllm_gen_helix(

tests/integration/test_lists/qa/llm_function_core.txt

@@ -1,24 +1,10 @@
 accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype[False]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1tp1cp4]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp2tp1cp2]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype[mtp_nextn=0-overlap_scheduler=False]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype[mtp_nextn=0-overlap_scheduler=True]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype[mtp_nextn=2-overlap_scheduler=False]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype[mtp_nextn=2-overlap_scheduler=True]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp1cp4]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp2tp1cp2]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp1cp4]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp2tp1cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1tp1cp4]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp2tp1cp2]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_gen_only_sync[python]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_guided_decoding[llguidance-mtp_nextn=0]
 accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_guided_decoding[llguidance-mtp_nextn=2]
@@ -87,18 +73,8 @@ accuracy/test_disaggregated_serving.py::TestQwen3_30B_A3B::test_mixed_ctx_gen_mo
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype[False-False]
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype[False-True]
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype[True-True]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp1cp4]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp2tp1cp2]
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp1cp4]
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp2tp1cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1dp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1tp1cp4]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp1tp2cp2]
-accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[nccl-cudagraph:with_padding-pp2tp1cp2]
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_chunked_prefill
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_gen_first
 accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_gen_first_kv_cache_v1
@@ -870,7 +846,6 @@ disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_co
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_conditional_v2[DeepSeek-V3-Lite-bf16]
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_empty_batch[DeepSeek-V3-Lite-bf16]
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_tllm_gen_helix[DeepSeek-V3-Lite-bf16-short_prompt]
-disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_tllm_gen_helix[DeepSeek-V3-Lite-bf16-long_prompt]
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_fp8_attention_dp[DeepSeek-V3-Lite-fp8]
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_fp8_attention_dp_gen_only[DeepSeek-V3-Lite-fp8]
 disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_fp8_attention_dp_one[DeepSeek-V3-Lite-fp8]

tests/integration/test_lists/test-db/l0_dgx_b200.yml

@@ -148,12 +148,8 @@ l0_dgx_b200:
       backend: pytorch
       orchestrator: mpi
   tests:
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp2tp1cp2] TIMEOUT (60)
   - accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1dp2cp2] TIMEOUT (60)
   - accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp2cp2] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v2-cudagraph:with_padding-pp1tp1cp4] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1tp1cp4] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1tp2cp2] TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput] TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput_mtp] TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput_bs8_mtp] TIMEOUT (60)
@@ -227,14 +223,7 @@ l0_dgx_b200:
       backend: pytorch
       orchestrator: mpi
   tests:
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp2cp2] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1tp1cp4] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp2tp1cp2] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestQwen3_8B::test_auto_dtype_with_helix[fifo_v1-cudagraph:with_padding-pp1dp2cp2] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp1dp2cp2] TIMEOUT (60)
-  - accuracy/test_disaggregated_serving.py::TestDeepSeekV32Exp::test_auto_dtype_with_helix[fifo-cudagraph:with_padding-pp2tp1cp2] TIMEOUT (60)
   - disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_tllm_gen_helix[DeepSeek-V3-Lite-bf16-short_prompt] TIMEOUT (60)
-  - disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_bf16_tllm_gen_helix[DeepSeek-V3-Lite-bf16-long_prompt] TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus_corner_case TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekV32::test_fp8_blockscale[baseline_fp8kv] TIMEOUT (60)
   - accuracy/test_llm_api_pytorch.py::TestDeepSeekV32::test_fp8_blockscale[latency] TIMEOUT (60)