[PATCH] D156579: InstCombine: Fold fneg (ldexp x, n) -> ldexp (fneg x), n

[Joshua Cranmer via Phabricator via llvm-commits]

jcranmer-intel added inline comments.


================
Comment at: llvm/test/Transforms/InstCombine/fneg.ll:840
+; CHECK-NEXT:    ret float [[NEG]]
+;
+  %ldexp = call float @llvm.ldexp.f32.i32(float %x, i32 %n)
----------------
The optimization doesn't seem to be kicking in for these tests?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D156579/new/

https://reviews.llvm.org/D156579