From ba809266818cfb9e63bcb34d79fdd77af6e308fe Mon Sep 17 00:00:00 2001 From: rasmith Date: Thu, 11 Dec 2025 19:02:19 -0600 Subject: [PATCH] [CI/Build][AMD] Skip test_cutlass_w4a8_moe tests on ROCm sine they require cutlass_pack_scale_fp8 (#30508) Signed-off-by: Randall Smith Signed-off-by: Michael Goin Signed-off-by: mgoin Co-authored-by: Randall Smith Co-authored-by: Michael Goin Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> --- tests/kernels/quantization/test_cutlass_w4a8_moe.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tests/kernels/quantization/test_cutlass_w4a8_moe.py b/tests/kernels/quantization/test_cutlass_w4a8_moe.py index 3560402a29e90..a855f7333b617 100644 --- a/tests/kernels/quantization/test_cutlass_w4a8_moe.py +++ b/tests/kernels/quantization/test_cutlass_w4a8_moe.py @@ -18,7 +18,9 @@ from vllm.model_executor.layers.quantization.utils.quant_utils import ( from vllm.platforms import current_platform from vllm.scalar_type import ScalarType, scalar_types -IS_SUPPORTED_BY_GPU = current_platform.get_device_capability()[0] >= 9 +IS_SUPPORTED_BY_GPU = ( + current_platform.is_cuda() and current_platform.get_device_capability()[0] >= 9 +) def to_fp8(tensor: torch.Tensor) -> torch.Tensor: