[PATCH] D101062: [AArch64][SVE] Better utilisation of unpredicated forms of arithmetic intrinsics

Bradley Smith via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Thu May 6 03:47:09 PDT 2021


bsmith added inline comments.


================
Comment at: llvm/lib/Target/AArch64/AArch64ISelLowering.cpp:17693
+bool AArch64TargetLowering::isAllActivePredicate(SDValue N) const {
+  return ::isAllActivePredicate(N);
+}
----------------
sdesmalen wrote:
> What is the value of implementing `isAllActivePredicate` in a separate, local function?
The issue here is that `convertMergedOpToPredOp` also needs to be able to use `isAllActivePredicate`, but `convertMergedOpToPredOp` is also local as well as things further up the call tree (and I couldn't find a sensible way to access AArch64TargetLowering from the local function, given the available state, as well as that just feeling wrong), a sensible alternative to having this 'wrapper' function would be to make `performIntrinsicCombine` a member function, which is fine, I just wanted to avoid changing too much when there is no precedent.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D101062/new/

https://reviews.llvm.org/D101062



More information about the llvm-commits mailing list