[PATCH] D28792: AMDGPU: Fold fneg into fminnum/fmaxnum

[Konstantin Zhuravlyov via Phabricator via llvm-commits]

kzhuravl added a comment.

LGTM with minor comment.



================
Comment at: lib/Target/AMDGPU/AMDGPUISelLowering.cpp:2991-2992
+  case ISD::FMINNUM: {
+    // fneg (fmaxnum x, y) -> fmaxnum (fneg x), (fneg y)
+    // fneg (fminnum x, y) -> fminnum (fneg x), (fneg y)
+    SDValue LHS = N0.getOperand(0);
----------------
I think this comments should be:
```
// fneg (fmaxnum x, y) -> fminnum (fneg x), (fneg y)
// fneg (fminnum x, y) -> fmaxnum (fneg x), (fneg y)
```


https://reviews.llvm.org/D28792