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

[via llvm-commits]

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

I have purposely kept the function arguments' names same as those in the comment at the  call-site/entry point of the optimization so that a reader can easily correlate things.
Still would you suggest to have comment?

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