2 files changed, 27 insertions, 1 deletions

diff --git a/llvm/test/Transforms/LoopVectorize/flags.ll b/llvm/test/Transforms/LoopVectorize/flags.ll
index a4ebb428488..21d09372d54 100644
--- a/llvm/test/Transforms/LoopVectorize/flags.ll
+++ b/llvm/test/Transforms/LoopVectorize/flags.ll
@@ -51,3 +51,29 @@ define i32 @flags2(i32 %n, i32* nocapture %A) nounwind uwtable ssp {
 ._crit_edge:                                      ; preds = %.lr.ph, %0
   ret i32 undef
 }
+
+; Make sure we copy fast math flags and use them for the final reduction.
+; CHECK-LABEL: fast_math
+; CHECK: load <4 x float>
+; CHECK: fadd fast <4 x float>
+; CHECK: br
+; CHECK: fadd fast <4 x float>
+; CHECK: fadd fast <4 x float>
+define float @fast_math(float* noalias %s) {
+entry:
+  br label %for.body
+
+for.body:
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %q.04 = phi float [ 0.000000e+00, %entry ], [ %add, %for.body ]
+  %arrayidx = getelementptr inbounds float* %s, i64 %indvars.iv
+  %0 = load float* %arrayidx, align 4
+  %add = fadd fast float %q.04, %0
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond = icmp eq i64 %indvars.iv.next, 256
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:
+  %add.lcssa = phi float [ %add, %for.body ]
+  ret float %add.lcssa
+}
diff --git a/llvm/test/Transforms/LoopVectorize/float-reduction.ll b/llvm/test/Transforms/LoopVectorize/float-reduction.ll
index c45098dd2c3..0dfbab07279 100644
--- a/llvm/test/Transforms/LoopVectorize/float-reduction.ll
+++ b/llvm/test/Transforms/LoopVectorize/float-reduction.ll
@@ -3,7 +3,7 @@
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 target triple = "x86_64-apple-macosx10.8.0"
 ;CHECK-LABEL: @foo(
-;CHECK: fadd <4 x float>
+;CHECK: fadd fast <4 x float>
 ;CHECK: ret
 define float @foo(float* nocapture %A, i32* nocapture %n) nounwind uwtable readonly ssp {
 entry: