[PATCH] D146035: [AArch64] Add FP16 broadcast and transpose costs

Tue Mar 14 05:05:59 PDT 2023

dmgreen created this revision.
dmgreen added reviewers: SjoerdMeijer, samtebbs, jaykang10.
Herald added subscribers: hiraditya, kristof.beyls.
Herald added a project: All.
dmgreen requested review of this revision.
Herald added a project: LLVM.

The FP16 broadcast and transpose can always use the same instructions as are used for i16 vectors, with or without +fullfp16. This fills in some extra costs to make sure we get it right.

https://reviews.llvm.org/D146035

Files:
  llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
  llvm/test/Analysis/CostModel/AArch64/shuffle-load.ll
  llvm/test/Analysis/CostModel/AArch64/shuffle-transpose.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D146035.505036.patch
Type: text/x-patch
Size: 7284 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230314/8bf5d9de/attachment.bin>