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 14, 05:30 UTC
[Log] Optimize DeepGEMM Missing Log (#26106)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Oct 2, 2025 at 22:02 UTC
36ce76c6321eff9467ee6ff4f2689b84eb14b48d
1 file changed
+2
-2
vllm/utils/deep_gemm.py
+2
-2
@@ -68,8 +68,8 @@ def is_deep_gemm_e8m0_used() -> bool:
68
def _missing(*_: Any, **__: Any) -> NoReturn:
69
"""Placeholder for unavailable DeepGEMM backend."""
70
raise RuntimeError(
71
- "DeepGEMM backend is not available. Please install the `deep_gemm` "
72
- "package to enable FP8 kernels.")
71
+ "DeepGEMM backend is not available or outdated. Please install or "
72
+ "update the `deep_gemm` to a newer version to enable FP8 kernels.")
73
74
75
_fp8_gemm_nt_impl: Callable[..., Any] | None = None