[all-commits] [llvm/llvm-project] 560b70: [X86][AVX] Cleanup AVX2 vector integer truncation ...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 560b709abedc76d7944be38d2f38b7bc8d40e32d
      https://github.com/llvm/llvm-project/commit/560b709abedc76d7944be38d2f38b7bc8d40e32d
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2021-05-18 (Tue, 18 May 2021)

  Changed paths:
    M llvm/lib/Target/X86/X86TargetTransformInfo.cpp
    M llvm/test/Analysis/CostModel/X86/cast.ll
    M llvm/test/Analysis/CostModel/X86/trunc.ll

  Log Message:
  -----------
  [X86][AVX] Cleanup AVX2 vector integer truncation costs

Noticed while investigating PR50364, the truncation costs for v4i64->v4i16/v4i8 and v8i32->v8i8 were way too optimistic for a shuffle sequence that usually matches the AVX1 codegen (they matched AVX512 numbers which have actual truncation instructions!).