[TensorExpr] Add a boilerplate pass for future TensorExpr fusion pass. (#33464)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/33464

I added a python-exposed knob to register this pass in custom passes pipeline. If the knob is not used, the pass is not registered and thus not run at all.

Differential Revision: D19958217

Test Plan: Imported from OSS

Pulled By: ZolotukhinM

fbshipit-source-id: fecdd98567fcda069fbdf8995c796899a3dbfa5c
This commit is contained in:
Mikhail Zolotukhin
2020-02-24 18:45:55 -08:00
committed by Facebook Github Bot
parent 9278196d89
commit bf00b4d305
5 changed files with 227 additions and 0 deletions

View File

@ -411,6 +411,7 @@ if (NOT INTERN_BUILD_MOBILE OR NOT BUILD_CAFFE2_MOBILE)
${TORCH_SRC_DIR}/csrc/jit/passes/requires_grad_analysis.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/specialize_autogradzero.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/subgraph_rewrite.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/tensorexpr_fuser.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/python_print.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/utils/subgraph_utils.cpp
${TORCH_SRC_DIR}/csrc/jit/passes/utils/check_alias_annotation.cpp