1; RUN: opt < %s -loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -dce -instcombine -S | FileCheck %s 2 3target 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" 4target triple = "x86_64-apple-macosx10.9.0" 5 6; Make sure we vectorize this loop: 7; int foo(float *a, float *b, int n) { 8; for (int i=0; i<n; ++i) 9; a[i] = b[i] * 3; 10; } 11 12;CHECK-LABEL: define i32 @foo 13;CHECK: for.body.preheader: 14;CHECK: br i1 %cmp.zero, label %scalar.ph, label %vector.memcheck, !dbg [[BODY_LOC:![0-9]+]] 15;CHECK: vector.memcheck: 16;CHECK: br i1 %memcheck.conflict, label %scalar.ph, label %vector.ph, !dbg [[BODY_LOC]] 17;CHECK: load <4 x float> 18define i32 @foo(float* nocapture %a, float* nocapture %b, i32 %n) nounwind uwtable ssp { 19entry: 20 %cmp6 = icmp sgt i32 %n, 0, !dbg !6 21 br i1 %cmp6, label %for.body, label %for.end, !dbg !6 22 23for.body: ; preds = %entry, %for.body 24 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ], !dbg !7 25 %arrayidx = getelementptr inbounds float, float* %b, i64 %indvars.iv, !dbg !7 26 %0 = load float, float* %arrayidx, align 4, !dbg !7 27 %mul = fmul float %0, 3.000000e+00, !dbg !7 28 %arrayidx2 = getelementptr inbounds float, float* %a, i64 %indvars.iv, !dbg !7 29 store float %mul, float* %arrayidx2, align 4, !dbg !7 30 %indvars.iv.next = add i64 %indvars.iv, 1, !dbg !7 31 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !7 32 %exitcond = icmp eq i32 %lftr.wideiv, %n, !dbg !7 33 br i1 %exitcond, label %for.end, label %for.body, !dbg !7 34 35for.end: ; preds = %for.body, %entry 36 ret i32 undef, !dbg !8 37} 38 39; Make sure that we try to vectorize loops with a runtime check if the 40; dependency check fails. 41 42; CHECK-LABEL: test_runtime_check 43; CHECK: <4 x float> 44define void @test_runtime_check(float* %a, float %b, i64 %offset, i64 %offset2, i64 %n) { 45entry: 46 br label %for.body 47 48for.body: 49 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] 50 %ind.sum = add i64 %iv, %offset 51 %arr.idx = getelementptr inbounds float, float* %a, i64 %ind.sum 52 %l1 = load float, float* %arr.idx, align 4 53 %ind.sum2 = add i64 %iv, %offset2 54 %arr.idx2 = getelementptr inbounds float, float* %a, i64 %ind.sum2 55 %l2 = load float, float* %arr.idx2, align 4 56 %m = fmul fast float %b, %l2 57 %ad = fadd fast float %l1, %m 58 store float %ad, float* %arr.idx, align 4 59 %iv.next = add nuw nsw i64 %iv, 1 60 %exitcond = icmp eq i64 %iv.next, %n 61 br i1 %exitcond, label %loopexit, label %for.body 62 63loopexit: 64 ret void 65} 66 67; CHECK: [[BODY_LOC]] = !DILocation(line: 101, column: 1, scope: !{{.*}}) 68 69!llvm.module.flags = !{!0, !1} 70!llvm.dbg.cu = !{!9} 71!0 = !{i32 2, !"Dwarf Version", i32 4} 72!1 = !{i32 2, !"Debug Info Version", i32 3} 73 74!2 = !{} 75!3 = !DISubroutineType(types: !2) 76!4 = !DIFile(filename: "test.cpp", directory: "/tmp") 77!5 = distinct !DISubprogram(name: "foo", scope: !4, file: !4, line: 99, type: !3, isLocal: false, isDefinition: true, scopeLine: 100, flags: DIFlagPrototyped, isOptimized: false, unit: !9, variables: !2) 78!6 = !DILocation(line: 100, column: 1, scope: !5) 79!7 = !DILocation(line: 101, column: 1, scope: !5) 80!8 = !DILocation(line: 102, column: 1, scope: !5) 81!9 = distinct !DICompileUnit(language: DW_LANG_C99, producer: "clang", 82 file: !10, 83 isOptimized: true, flags: "-O2", 84 splitDebugFilename: "abc.debug", emissionKind: 2) 85!10 = !DIFile(filename: "path/to/file", directory: "/path/to/dir") 86!11 = !{i32 2, !"Debug Info Version", i32 3} 87