Home | History | Annotate | Download | only in LoopVectorize
      1 ; RUN: opt < %s  -loop-vectorize -force-vector-unroll=4 -force-vector-width=4 -debug-only=loop-vectorize -stats -S 2>&1 | FileCheck %s
      2 ; REQUIRES: asserts
      3 
      4 ;
      5 ; We have 2 loops, one of them is vectorizable and the second one is not.
      6 ;
      7 
      8 ; CHECK: 2 loop-vectorize               - Number of loops analyzed for vectorization
      9 ; CHECK: 1 loop-vectorize               - Number of loops vectorized
     10 
     11 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"
     12 target triple = "x86_64-unknown-linux-gnu"
     13 
     14 define void @vectorized(float* nocapture %a, i64 %size) {
     15 entry:
     16   %cmp1 = icmp sgt i64 %size, 0
     17   br i1 %cmp1, label %for.header, label %for.end
     18 
     19 for.header:
     20   %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
     21   %cmp2 = icmp sgt i64 %indvars.iv, %size
     22   br i1 %cmp2, label %for.end, label %for.body
     23 
     24 for.body:
     25 
     26   %arrayidx = getelementptr inbounds float* %a, i64 %indvars.iv
     27   %0 = load float* %arrayidx, align 4
     28   %mul = fmul float %0, %0
     29   store float %mul, float* %arrayidx, align 4
     30 
     31   %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
     32   br label %for.header
     33 
     34 for.end:
     35   ret void
     36 }
     37 
     38 define void @not_vectorized(float* nocapture %a, i64 %size) {
     39 entry:
     40   %cmp1 = icmp sgt i64 %size, 0
     41   br i1 %cmp1, label %for.header, label %for.end
     42 
     43 for.header:
     44   %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
     45   %cmp2 = icmp sgt i64 %indvars.iv, %size
     46   br i1 %cmp2, label %for.end, label %for.body
     47 
     48 for.body:
     49 
     50   %0 = add nsw i64 %indvars.iv, -5
     51   %arrayidx = getelementptr inbounds float* %a, i64 %0
     52   %1 = load float* %arrayidx, align 4
     53   %2 = add nsw i64 %indvars.iv, 2
     54   %arrayidx2 = getelementptr inbounds float* %a, i64 %2
     55   %3 = load float* %arrayidx2, align 4
     56   %mul = fmul float %1, %3
     57   %arrayidx4 = getelementptr inbounds float* %a, i64 %indvars.iv
     58   store float %mul, float* %arrayidx4, align 4
     59 
     60   %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
     61   br label %for.header
     62 
     63 for.end:
     64   ret void
     65 }