[PATCH] D93594: [X86] Pass to transform amx intrinsics to scalar operation.

[LuoYuanke via Phabricator via llvm-commits]

LuoYuanke created this revision.
Herald added subscribers: nikic, pengfei, hiraditya, mgorny.
LuoYuanke requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

This pass is only enabled with -O0. With -O0, the def of shape to amx
intrinsics is near the amx intrinsics code. We are not able to find a
point which post-dominate all the shape and dominate all amx intrinsics.
To decouple the dependency of the shape, we transform amx intrinsics
to scalar operation, so that compiling doesn't fail. In long term, we
should improve fast register allocation to allocate amx register.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D93594

Files:
  llvm/include/llvm/CodeGen/Passes.h
  llvm/lib/Target/X86/CMakeLists.txt
  llvm/lib/Target/X86/X86.h
  llvm/lib/Target/X86/X86LowerAMXIntrinsics.cpp
  llvm/lib/Target/X86/X86TargetMachine.cpp
  llvm/test/CodeGen/X86/AMX/amx-low-intrinsics.ll
  llvm/test/CodeGen/X86/O0-pipeline.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D93594.312970.patch
Type: text/x-patch
Size: 25299 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20201220/9f0d827e/attachment.bin>