[PATCH] D103169: [FPEnv][InstSimplify] Constrained FP support for NaN

[Kevin P. Neal via Phabricator via llvm-commits]

kpn added a comment.

Phab just lost a comment of mine. Weird.



================
Comment at: llvm/test/Transforms/InstSimplify/X86/fp-nan-strictfp.ll:305
+
+define float @frem_nan_op0_upward(float %x) {
+; CHECK-LABEL: @frem_nan_op0_upward(
----------------
sepavloff wrote:
> Calculation of remainder does not depend on rounding mode, so this test is identical to the next.
Changing the rounding mode means we're no longer in the default FP environment. This then changes the codepath through SimplifyFRemInst(). That's why I put this test here.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D103169/new/

https://reviews.llvm.org/D103169