[all-commits] [llvm/llvm-project] fcb834: [LoongArch] Add definitions and feature 'frecipe' ...

[Ami-zhang via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fcb8342a219ada8ec641790a4c8a9f969d7d64ee
      https://github.com/llvm/llvm-project/commit/fcb8342a219ada8ec641790a4c8a9f969d7d64ee
  Author: Ami-zhang <zhanglimin at loongson.cn>
  Date:   2024-01-23 (Tue, 23 Jan 2024)

  Changed paths:
    M clang/include/clang/Basic/BuiltinsLoongArchBase.def
    M clang/include/clang/Basic/BuiltinsLoongArchLASX.def
    M clang/include/clang/Basic/BuiltinsLoongArchLSX.def
    M clang/lib/Headers/larchintrin.h
    M clang/lib/Headers/lasxintrin.h
    M clang/lib/Headers/lsxintrin.h
    A clang/test/CodeGen/LoongArch/builtin-dbl-approximate.c
    A clang/test/CodeGen/LoongArch/builtin-flt-approximate.c
    M clang/test/CodeGen/LoongArch/intrinsic-la64-error.c
    A clang/test/CodeGen/LoongArch/lasx/builtin-approximate-alias.c
    A clang/test/CodeGen/LoongArch/lasx/builtin-approximate.c
    A clang/test/CodeGen/LoongArch/lsx/builtin-approximate-alias.c
    A clang/test/CodeGen/LoongArch/lsx/builtin-approximate.c
    M llvm/include/llvm/IR/IntrinsicsLoongArch.td
    M llvm/lib/Target/LoongArch/LoongArch.td
    M llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td
    M llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td
    M llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
    M llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
    M llvm/lib/Target/LoongArch/LoongArchSubtarget.h
    A llvm/test/CodeGen/LoongArch/intrinsic-frecipe-dbl.ll
    A llvm/test/CodeGen/LoongArch/intrinsic-frecipe-flt.ll
    A llvm/test/CodeGen/LoongArch/lasx/intrinsic-frecipe.ll
    A llvm/test/CodeGen/LoongArch/lasx/intrinsic-frsqrte.ll
    A llvm/test/CodeGen/LoongArch/lsx/intrinsic-frecipe.ll
    A llvm/test/CodeGen/LoongArch/lsx/intrinsic-frsqrte.ll

  Log Message:
  -----------
  [LoongArch] Add definitions and feature 'frecipe' for FP approximation intrinsics/builtins (#78962)

This PR adds definitions and 'frecipe' feature for FP approximation
intrinsics/builtins. In additions, this adds and complements relative
testcases.