[PATCH] D131964: [AArch64] Add support to loop vectorization for non temporal loads

[Florian Hahn via Phabricator via llvm-commits]

fhahn added inline comments.


================
Comment at: llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h:326
+      unsigned NumElements = DataTypeTy->getNumElements();
+      unsigned EltSize = DataTypeTy->getElementType()->getScalarSizeInBits();
+      return NumElements > 1 && isPowerOf2_64(NumElements) && EltSize >= 8 &&
----------------
Is there much difference between the load and store version? Could they just share the same code?

I think this would also require support for generating `LDNP` for types smaller than 256 bits. @zjaffal  is currently looking into this.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D131964/new/

https://reviews.llvm.org/D131964