[PATCH] D142594: [AArch64] Eliminating the use of integer unit in moving from a Neon scalar result of a uaddlv to a Neon vector

[NILANJANA BASU via Phabricator via llvm-commits]

nilanjana_basu added a comment.

Thank you everyone for the feedback. I am currently working on the post-isel peephole optimization since that can capture the more generic pattern of

%y:gpr = COPY %x:fpr
%z:fpr = INSvi..gpr .., a, %y:gpr

& replace it with
%z:fpr = INSvi..lane .., a, %x:fpr, 0

which should apply to a larger class of code than vector insertion of the uaddlv intrinsic result.
I'll update the patch & problem description once I have a fully working solution.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142594/new/

https://reviews.llvm.org/D142594