[PATCH] D26575: Fix for lost FastMathFlags in SLPVectorizer (intrinsic calls)

[Michael Zolotukhin via llvm-commits]

mzolotukhin accepted this revision.
mzolotukhin added a comment.
This revision is now accepted and ready to land.

LGTM, thanks for following-up!

Michael



================
Comment at: llvm/test/Transforms/SLPVectorizer/X86/propagate_ir_flags.ll:413
+  %call1 = tail call fast double @llvm.fabs.f64(double %load1) nounwind readnone
+  %call2 = tail call fast double @llvm.fabs.f64(double %load1) nounwind readnone
+
----------------
Should it be `%load2`?


================
Comment at: llvm/test/Transforms/SLPVectorizer/X86/propagate_ir_flags.ll:431
+  %call1 = tail call fast double @llvm.fabs.f64(double %load1) nounwind readnone
+  %call2 = tail call double @llvm.fabs.f64(double %load1) nounwind readnone
+
----------------
The same.


https://reviews.llvm.org/D26575