[llvm] [Vectorize] Vectorization for __builtin_prefetch (PR #66160)
via llvm-commits
llvm-commits at lists.llvm.org
Wed Dec 13 18:44:03 PST 2023
================
@@ -539,6 +539,27 @@ CallInst *IRBuilderBase::CreateMaskedStore(Value *Val, Value *Ptr,
return CreateMaskedIntrinsic(Intrinsic::masked_store, Ops, OverloadedTypes);
}
+/// Create a call to a Masked Prefetch intrinsic.
+/// \p Ptr - base pointer for the load
+/// \p ElemSize - element size for memory address generation
+/// \p Mask - vector of booleans which indicates what vector lanes should
+/// be accessed in memory
+/// \p RW - Read or Write
+/// \p Locality - Cache Level
+/// \p Name - name of the result variable
+CallInst *IRBuilderBase::CreateMaskedPrefetch(Value *Ptr,
+ Value *ElemSize,
+ Value *Mask, Value *RW,
+ Value *Locality,
+ const Twine &Name) {
+ auto *PtrTy = cast<PointerType>(Ptr->getType());
+ assert(Mask && "Mask should not be all-ones (null)");
----------------
m-saito-fj wrote:
Thanks for your comment.
I had just done the same with masked load/masked store.
I think that if `mask` is `null`, then mask can be considered unnecessary. So I think there is no problem with `all ones`.
But, I think it would be confusing with that message as it was, so I fixed it.
[assert message](https://github.com/llvm/llvm-project/pull/66160/files#diff-7b9ae5c1774880299e98d0bd713f6d0b881a6f70ad31ce5603ea6275fecf8a42R556).
The comment is outdated because I force pushed it. Please see link above.
https://github.com/llvm/llvm-project/pull/66160
More information about the llvm-commits
mailing list