Skip to content

GitLab

Unverified Commit 56dcf4e7 authored by Wentao Ye's avatar Wentao Ye Committed by GitHub
Browse files

[Bug] Fix DeepGEMM Env Control (#23591) 


Signed-off-by: default avataryewentao256 <zhyanwentao@126.com>
parent ae067888
Hide whitespace changes
Inline Side-by-side
Showing with 4 additions and 7 deletions
vllm/utils/deep_gemm.py
View file @ 56dcf4e7
...@@ -27,7 +27,7 @@ def is_deep_gemm_supported() -> bool: ...@@ -27,7 +27,7 @@ def is_deep_gemm_supported() -> bool:
is_supported_arch = current_platform.is_cuda() and ( is_supported_arch = current_platform.is_cuda() and (
current_platform.is_device_capability(90) current_platform.is_device_capability(90)
or current_platform.is_device_capability(100)) or current_platform.is_device_capability(100))
return has_deep_gemm() and is_supported_arch return envs.VLLM_USE_DEEP_GEMM and has_deep_gemm() and is_supported_arch
@functools.cache @functools.cache
...@@ -35,12 +35,9 @@ def is_blackwell_deep_gemm_e8m0_used() -> bool: ...@@ -35,12 +35,9 @@ def is_blackwell_deep_gemm_e8m0_used() -> bool:
"""Return ``True`` if vLLM is configured to use DeepGEMM " """Return ``True`` if vLLM is configured to use DeepGEMM "
"E8M0 scale on a Blackwell-class GPU. "E8M0 scale on a Blackwell-class GPU.
""" """
if not (envs.VLLM_USE_DEEP_GEMM): if not is_deep_gemm_supported():
logger.debug_once("DeepGEMM E8M0 disabled: VLLM_USE_DEEP_GEMM=0.") logger.debug_once(
return False "DeepGEMM E8M0 disabled: DeepGEMM not supported on this system.")
if not has_deep_gemm():
logger.debug_once("DeepGEMM E8M0 disabled: DeepGEMM backend missing.")
return False return False
if not envs.VLLM_USE_DEEP_GEMM_E8M0: if not envs.VLLM_USE_DEEP_GEMM_E8M0:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment