From a0c093c4c9fe28d29fe121511b6273c583a6af96 Mon Sep 17 00:00:00 2001 From: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> Date: Wed, 20 Aug 2025 07:35:57 +0000 Subject: [PATCH 1/4] [TRTLLM-7363][test] Add 8-GPU test cases for RTX6000 Signed-off-by: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> --- tests/integration/test_lists/qa/llm_function_rtx6kd.txt | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt index b3d14c393bb..0fdd8955d0d 100644 --- a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt +++ b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt @@ -26,6 +26,7 @@ accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_cutl accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_cutlass-torch_compile=True] accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_trtllm-torch_compile=False] accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_trtllm-torch_compile=True] +accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput_tp8] test_e2e.py::test_ptp_quickstart_advanced_mixed_precision test_e2e.py::test_ptp_quickstart_advanced[Llama3.1-8B-BF16-llama-3.1-model/Meta-Llama-3.1-8B] test_e2e.py::test_ptp_quickstart_advanced[Llama3.1-8B-FP8-llama-3.1-model/Llama-3.1-8B-Instruct-FP8] @@ -39,3 +40,4 @@ test_e2e.py::test_ptp_quickstart_advanced[Mixtral-8x7B-FP8-Mixtral-8x7B-Instruct test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Llama3.1-70B-BF16-llama-3.1-model/Meta-Llama-3.1-70B] test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Nemotron-Super-49B-v1-BF16-nemotron-nas/Llama-3_3-Nemotron-Super-49B-v1] test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Mixtral-8x7B-BF16-Mixtral-8x7B-Instruct-v0.1] +test_e2e.py::test_ptp_quickstart_advanced_deepseek_r1_8gpus[DeepSeek-R1-DeepSeek-R1/DeepSeek-R1] From d0c288ed9f133a6bfbb963d19838be89f0851553 Mon Sep 17 00:00:00 2001 From: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> Date: Wed, 20 Aug 2025 07:59:03 +0000 Subject: [PATCH 2/4] test ds-r1 fp4 Signed-off-by: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> --- tests/integration/defs/test_e2e.py | 3 +++ tests/integration/test_lists/qa/llm_function_rtx6kd.txt | 2 +- 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/tests/integration/defs/test_e2e.py b/tests/integration/defs/test_e2e.py index 478a89f5a26..0375ddc6a2e 100644 --- a/tests/integration/defs/test_e2e.py +++ b/tests/integration/defs/test_e2e.py @@ -2071,6 +2071,9 @@ def test_ptp_quickstart_advanced_deepseek_v3_lite_4gpus_adp_balance( @pytest.mark.parametrize("model_name,model_path", [ pytest.param( 'DeepSeek-R1', 'DeepSeek-R1/DeepSeek-R1', marks=skip_pre_hopper), + pytest.param('DeepSeek-R1-0528-FP4', + 'DeepSeek-R1/DeepSeek-R1-0528-FP4', + marks=skip_pre_blackwell) ]) def test_ptp_quickstart_advanced_deepseek_r1_8gpus(llm_root, llm_venv, model_name, model_path): diff --git a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt index 0fdd8955d0d..5d797a292c7 100644 --- a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt +++ b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt @@ -40,4 +40,4 @@ test_e2e.py::test_ptp_quickstart_advanced[Mixtral-8x7B-FP8-Mixtral-8x7B-Instruct test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Llama3.1-70B-BF16-llama-3.1-model/Meta-Llama-3.1-70B] test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Nemotron-Super-49B-v1-BF16-nemotron-nas/Llama-3_3-Nemotron-Super-49B-v1] test_e2e.py::test_ptp_quickstart_advanced_2gpus_sm120[Mixtral-8x7B-BF16-Mixtral-8x7B-Instruct-v0.1] -test_e2e.py::test_ptp_quickstart_advanced_deepseek_r1_8gpus[DeepSeek-R1-DeepSeek-R1/DeepSeek-R1] +test_e2e.py::test_ptp_quickstart_advanced_deepseek_r1_8gpus[DeepSeek-R1-0528-FP4-DeepSeek-R1/DeepSeek-R1-0528-FP4] From 85f4e471765837295454359e6e80137204899d2f Mon Sep 17 00:00:00 2001 From: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> Date: Thu, 21 Aug 2025 08:07:20 +0000 Subject: [PATCH 3/4] add Qwen3_235B_A22B test cases Signed-off-by: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> --- tests/integration/test_lists/qa/llm_function_rtx6kd.txt | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt index 5d797a292c7..69f0787908a 100644 --- a/tests/integration/test_lists/qa/llm_function_rtx6kd.txt +++ b/tests/integration/test_lists/qa/llm_function_rtx6kd.txt @@ -26,6 +26,9 @@ accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_cutl accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_cutlass-torch_compile=True] accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_trtllm-torch_compile=False] accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[latency_moe_trtllm-torch_compile=True] +accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_nvfp4[latency_moe_cutlass] +accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_nvfp4[latency_moe_trtllm] +accuracy/test_llm_api_pytorch.py::TestQwen3_235B_A22B::test_nvfp4[latency_moe_trtllm_eagle3] accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[throughput_tp8] test_e2e.py::test_ptp_quickstart_advanced_mixed_precision test_e2e.py::test_ptp_quickstart_advanced[Llama3.1-8B-BF16-llama-3.1-model/Meta-Llama-3.1-8B] From e5afd21318a967caa041b1adb007c8c13639ff15 Mon Sep 17 00:00:00 2001 From: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> Date: Wed, 3 Sep 2025 07:32:56 +0000 Subject: [PATCH 4/4] add missing comma Signed-off-by: Stanley Sun <190317771+StanleySun639@users.noreply.github.com> --- tests/integration/defs/test_e2e.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/integration/defs/test_e2e.py b/tests/integration/defs/test_e2e.py index 0375ddc6a2e..152676bb9ab 100644 --- a/tests/integration/defs/test_e2e.py +++ b/tests/integration/defs/test_e2e.py @@ -2073,7 +2073,7 @@ def test_ptp_quickstart_advanced_deepseek_v3_lite_4gpus_adp_balance( 'DeepSeek-R1', 'DeepSeek-R1/DeepSeek-R1', marks=skip_pre_hopper), pytest.param('DeepSeek-R1-0528-FP4', 'DeepSeek-R1/DeepSeek-R1-0528-FP4', - marks=skip_pre_blackwell) + marks=skip_pre_blackwell), ]) def test_ptp_quickstart_advanced_deepseek_r1_8gpus(llm_root, llm_venv, model_name, model_path):