[TensorExpr] Add a fuser pass based on tensor expressions. (#34226)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34226

LLVM and Cuda backends are added in subsequent PRs, so at this point the fuser is pretty useless, but it still can be tested and its logic is not going to change with addition of the codegens.

Differential Revision: D20251838

Test Plan: Imported from OSS

Pulled By: ZolotukhinM

fbshipit-source-id: 82b0d221fa89904ed526689d02a6c7676a8ce8de
This commit is contained in:
Mikhail Zolotukhin
2020-03-16 11:38:29 -07:00
committed by Facebook GitHub Bot
parent e31d462e92
commit 42b2c8c65d
12 changed files with 2745 additions and 35 deletions

View File

@ -481,6 +481,7 @@ if (NOT INTERN_BUILD_MOBILE OR NOT BUILD_CAFFE2_MOBILE)
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/ir_mutator.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/ir_printer.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/ir_visitor.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/kernel.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/mem_arena.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/schedule.cpp
${TORCH_SRC_DIR}/csrc/jit/tensorexpr/tensor.cpp