{ROCm]: gpt-oss fusion/padding fixes (#38043)
Signed-off-by: Andreas Karatzas <akaratza@amd.com> Signed-off-by: Rohan138 <rohanpotdar138@gmail.com> Signed-off-by: Rohan Potdar <66227218+Rohan138@users.noreply.github.com> Co-authored-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
co-authored by
Andreas Karatzas
parent
21d2b53f88
commit
0e9358c11d
+1
-3
@@ -152,13 +152,11 @@ def enable_rope_kvcache_fusion(cfg: "VllmConfig") -> bool:
|
||||
|
||||
|
||||
def enable_norm_pad_fusion(cfg: "VllmConfig") -> bool:
|
||||
"""Enable if using AITER RMSNorm and AITER Triton GEMMs
|
||||
and hidden size is 2880 i.e. gpt-oss; otherwise Inductor handles fusion."""
|
||||
"""Enable if using AITER RMSNorm and hidden size is 2880 i.e. gpt-oss."""
|
||||
from vllm._aiter_ops import rocm_aiter_ops
|
||||
|
||||
return (
|
||||
rocm_aiter_ops.is_rmsnorm_enabled()
|
||||
and not rocm_aiter_ops.is_triton_gemm_enabled()
|
||||
and cfg.model_config is not None
|
||||
and cfg.model_config.get_hidden_size() == 2880
|
||||
)
|
||||
|
||||
Reference in New Issue
Block a user