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

[Joshua Cranmer via llvm-commits]

================
@@ -626,6 +626,100 @@ Instruction *InstCombinerImpl::foldPowiReassoc(BinaryOperator &I) {
   return nullptr;
 }
 
+static bool isFSqrtDivToFMulLegal(Instruction *X,
----------------
jcranmer-intel wrote:

Is it too much to ask for a comment on this method explaining what relation X, R1, and R2 are supposed to have to each other? It does make it hard to follow without one.

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