OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

Fix the AOTI compile failure with ARM CPU for Meta internal (#147204)

Browse Source 
Summary: Fix the AOTI compile failure with ARM CPU for Meta internal

Differential Revision: D69642211

Pull Request resolved: https://github.com/pytorch/pytorch/pull/147204
Approved by: https://github.com/houseroad
This commit is contained in:
Huamin Li 2025-02-18 17:54:34 +00:00 committed by PyTorch MergeBot
parent 5d675de754
commit dd2a943e14
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
torch/_inductor/cpu_vec_isa.py
View File

@ -160,6 +160,8 @@ class VecNEON(VecISA):
_dtype_nelements = {torch.float: 4, torch.bfloat16: 8, torch.float16: 8}
def __str__(self) -> str:
if config.is_fbcode():
return "neon"
return "asimd" # detects the presence of advanced SIMD on armv8-a kernels
__hash__: Callable[[VecISA], Any] = VecISA.__hash__
@ -178,6 +180,8 @@ class VecSVE256(VecISA):
_dtype_nelements = {torch.float: 8, torch.bfloat16: 16, torch.float16: 16}
def __str__(self) -> str:
if config.is_fbcode():
return "neon"
return "asimd"
__hash__: Callable[[VecISA], Any] = VecISA.__hash__