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

[Sushant Gokhale via llvm-commits]

================
@@ -1888,6 +2031,24 @@ Instruction *InstCombinerImpl::visitFDiv(BinaryOperator &I) {
     return R;
 
   Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);
+
+  // Convert
+  // x = 1.0/sqrt(a)
+  // r1 = x * x;
+  // r2 = a/sqrt(a);
+  //
+  // TO
+  //
+  // r1 = 1/a
+  // r2 = sqrt(a)
+  // x = r1 * r2
+  SmallPtrSet<Instruction *, 2> R1, R2;
----------------
sushgokh wrote:

Had to change from SmallVector to SmallPtrSet since invoking the `users()` API somehow got me duplicate users

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