From 1b6a9829cae9c8bb795162d95659af312e074244 Mon Sep 17 00:00:00 2001 From: Bo Li <22713281+bobboli@users.noreply.github.com> Date: Thu, 13 Aug 2026 15:34:01 +0800 Subject: [PATCH 1/3] [None][test] unwaive flaky accuracy tests Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com> --- tests/integration/test_lists/waives.txt | 1 - 1 file changed, 1 deletion(-) diff --git a/tests/integration/test_lists/waives.txt b/tests/integration/test_lists/waives.txt index 6adff766d605..caf00cb61b93 100644 --- a/tests/integration/test_lists/waives.txt +++ b/tests/integration/test_lists/waives.txt @@ -234,7 +234,6 @@ full:H100/accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mt full:H100/accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=True-enable_chunked_prefill=True-v2_kv_cache=False] SKIP (https://nvbugs/6662724) full:H100/accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=True-enable_chunked_prefill=True-v2_kv_cache=True] SKIP (https://nvbugs/6662724) full:H100/accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_dummy_load_format SKIP (https://nvbugs/6626640) -full:H100/accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B_Instruct_2507::test_skip_softmax_attention_4gpus[target_sparsity_0.9-fp8kv=False] SKIP (https://nvbugs/6523809) full:H100/accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_bf16[tp1-CUTLASS] SKIP (https://nvbugs/6273850) full:H100/accuracy/test_llm_api_pytorch_multimodal.py::TestExaone4_5_33B::test_auto_dtype[forced_chunked_prefill] SKIP (https://nvbugs/6597570) full:H100/accuracy/test_llm_api_pytorch_multimodal.py::TestExaone4_5_33B::test_auto_dtype[full_budget] SKIP (https://nvbugs/6597570) From a79c2205ff9a26da52271048800d2492bd672bfe Mon Sep 17 00:00:00 2001 From: Bo Li <22713281+bobboli@users.noreply.github.com> Date: Tue, 25 Aug 2026 15:39:32 +0800 Subject: [PATCH 2/3] [NVBUG/6458213][test] use TRTLLM MoE on Blackwell Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com> --- tests/integration/defs/accuracy/test_llm_api_pytorch.py | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/tests/integration/defs/accuracy/test_llm_api_pytorch.py b/tests/integration/defs/accuracy/test_llm_api_pytorch.py index 456f6895c7a9..555d74426887 100644 --- a/tests/integration/defs/accuracy/test_llm_api_pytorch.py +++ b/tests/integration/defs/accuracy/test_llm_api_pytorch.py @@ -5363,6 +5363,9 @@ def test_skip_softmax_attention(self, target_sparsity: float, with LLM(self.MODEL_PATH, attn_backend="TRTLLM", + moe_config=MoeConfig( + backend="TRTLLM" if get_sm_version() in (100, + 103) else "AUTO"), max_batch_size=256, max_num_tokens=100000, kv_cache_config=kv_cache_config, @@ -5399,6 +5402,9 @@ def test_skip_softmax_attention_4gpus(self, target_sparsity: float, with LLM(self.MODEL_PATH, attn_backend="TRTLLM", + moe_config=MoeConfig( + backend="TRTLLM" if get_sm_version() in (100, + 103) else "AUTO"), max_batch_size=256, max_num_tokens=100000, tensor_parallel_size=4, From ddcf249b5332e5f8af25c9bc501d0f24343900b1 Mon Sep 17 00:00:00 2001 From: Bo Li <22713281+bobboli@users.noreply.github.com> Date: Tue, 25 Aug 2026 16:14:49 +0800 Subject: [PATCH 3/3] [NVBUG/6606551][test] unwaive B200 skip-softmax test Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com> --- tests/integration/test_lists/waives.txt | 1 - 1 file changed, 1 deletion(-) diff --git a/tests/integration/test_lists/waives.txt b/tests/integration/test_lists/waives.txt index caf00cb61b93..1d42bad9ea95 100644 --- a/tests/integration/test_lists/waives.txt +++ b/tests/integration/test_lists/waives.txt @@ -75,7 +75,6 @@ accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_fp8[latency] SKIP (h accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_fp8[throughput_latency] SKIP (https://nvbugs/6177390) accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_nvfp4[latency_moe_trtllm] SKIP (https://nvbugs/6402009) accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[tep4_latency_moe_cutlass-torch_compile=True] SKIP (https://nvbugs/6561558) -accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B_Instruct_2507::test_skip_softmax_attention[target_sparsity_0.9-fp8kv=True] SKIP (https://nvbugs/6606551) accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_fp8[enable_block_reuse=True] SKIP (https://nvbugs/6625710) accuracy/test_llm_api_pytorch.py::TestQwen3_5_397B_A17B::test_nvfp4[adp4_cutedsl] SKIP (https://nvbugs/6535779) accuracy/test_llm_api_pytorch.py::TestQwen3_5_4B::test_dflash SKIP (https://nvbugs/6640134)