[Bugfix] Fix DeepGemm E8M0 accuracy degradation for Qwen3.5 FP8 on Blackwell (#38083)

Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
This commit is contained in:
Vadim Gimpelson
2026-03-26 01:21:47 -07:00
committed by GitHub
parent 71161e8b63
commit 52069012fe
10 changed files with 69 additions and 11 deletions
+18
View File
@@ -23,6 +23,24 @@ from vllm.platforms import current_platform
from vllm.utils.import_utils import has_deep_gemm
from vllm.utils.math_utils import cdiv
_DEEPGEMM_BLACKWELL_EXCLUDED_MODEL_TYPES: set[str] = {
"qwen3_5_text",
"qwen3_5_moe_text",
}
def should_auto_disable_deep_gemm(model_type: str | None) -> bool:
"""Check if DeepGemm should be auto-disabled for this model on Blackwell.
Returns True if the model is known to have accuracy degradation with
DeepGemm's E8M0 scale format on Blackwell GPUs (SM100+).
"""
if model_type is None:
return False
if not current_platform.is_device_capability_family(100):
return False
return model_type in _DEEPGEMM_BLACKWELL_EXCLUDED_MODEL_TYPES
class DeepGemmQuantScaleFMT(Enum):
# Float32 scales in Float32 tensor