A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 01:30 UTC
[CI][AMD] Skip test_permute_cols since the kernel is not used and not built for ROCm (#32444)
Signed-off-by: Randall Smith <ransmith@amd.com>
rasmith committed
Jan 16, 2026 at 02:22 UTC
6ca4f400d8b6c83a40af5c338b2ac55a8effdd76
1 file changed
+3
tests/kernels/core/test_permute_cols.py
+3
@@ -7,6 +7,9 @@ import torch
7
from tests.kernels.utils import opcheck
8
from vllm._custom_ops import permute_cols
9
10
+if not hasattr(torch.ops._C, "permute_cols"):
11
+ pytest.skip(reason="permute_cols is not supported on ROCm", allow_module_level=True)
12
+
13
14
@pytest.mark.parametrize("shape", [(1, 512), (544, 4096), (67, 8192)])
15
@pytest.mark.parametrize("dtype", [torch.bfloat16])