[PATCH] D18701: [ARM] Adding IEEE-754 SIMD detection to loop vectorizer

[James Greenhalgh via llvm-commits]

jgreenhalgh added a subscriber: jgreenhalgh.
jgreenhalgh added a comment.

Sorry to have missed this before commit.


================
Comment at: lib/Target/ARM/ARMTargetTransformInfo.h:58
@@ +57,3 @@
+  bool enablePotentiallyUnsafeFPVectorization() {
+    return ST->hasFPARMv8() || ST->isTargetDarwin();
+  }
----------------
What's the idea here? The FTZ behaviour doesn't change in AArch32 execution state on ARMv8-A, you're still unsafe. The full IEEE-compliant Advanced SIMD is in AArch64 execution state.


http://reviews.llvm.org/D18701