[PATCH] D78647: [MLIR] Lower GenericAtomicRMWOp to llvm.cmpxchg.

[Alexander Belyaev via Phabricator via llvm-commits]

pifon2a created this revision.
pifon2a added reviewers: herhut, flaub.
Herald added subscribers: llvm-commits, frgossen, grosul1, Joonsoo, liufengdb, aartbik, lucyrfox, mgester, arpith-jacob, nicolasvasilache, antiagainst, shauheen, jpienaar, rriddle, mehdi_amini, jfb.
Herald added a project: LLVM.
pifon2a added a comment.

What should we do with the lowering for AtomicRMWOp "minf", "maxf"? Should we lower it to GenericAtomicRMWOp and then to LLVM? Or maybe we should not use AtomicRMWOp for these cases at all? It can always be replaced by GenericAtomicRMWOp.


Lowering is pretty much a copy of AtomicRMWOp -> llvm.cmpxchg
pattern.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D78647

Files:
  mlir/include/mlir/Dialect/StandardOps/IR/Ops.td
  mlir/lib/Conversion/StandardToLLVM/StandardToLLVM.cpp
  mlir/test/Conversion/StandardToLLVM/convert-to-llvmir.mlir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D78647.259299.patch
Type: text/x-patch
Size: 7089 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200422/d21e9120/attachment.bin>