[TensorExpr] Add a way to define target triple/cpu/attrs for llvm codegen and turn on the AOT workflow. (#66527)
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/66527
Differential Revision:
D31593869
D31593869
Test Plan: Imported from OSS
Reviewed By: navahgar
Pulled By: ZolotukhinM
fbshipit-source-id: e7534c11fbcf0dab5f49d01d6053caf77b833ef0
Author
Mikhail Zolotukhin