[llvm] [InstCombine] Transform high latency, dependent FSQRT/FDIV into FMUL (PR #87474)

[via llvm-commits]

================
@@ -626,6 +626,127 @@ Instruction *InstCombinerImpl::foldPowiReassoc(BinaryOperator &I) {
   return nullptr;
 }
 
+bool isFSqrtDivToFMulLegal(Instruction *X, SmallSetVector<Instruction *, 2> &R1,
----------------
goldsteinn wrote:

static

https://github.com/llvm/llvm-project/pull/87474