[PATCH] D88278: [PowerPC] Add builtins for xvtdiv(dp|sp) and xvtsqrt(dp|sp).

[Qing Shan Zhang via Phabricator via llvm-commits]

steven.zhang added a comment.

I think some follow up is needed to optimize the code sequence for

  cr = vec_test_div
  if (rotate_and_mask(cr, 62))
    ...

For now, we will copy the cr to gpr, and shift it to 61-64 bit,then, extract the bit and then compare to 0. the shift is not needed. Please double confirm to see if the peephole handle it well.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D88278/new/

https://reviews.llvm.org/D88278