1 ; RUN: opt < %s -loop-vectorize -mtriple=x86_64-unknown-linux -S -pass-remarks='loop-vectorize' 2>&1 | FileCheck -check-prefix=VECTORIZED %s 2 ; RUN: opt < %s -loop-vectorize -force-vector-width=1 -force-vector-interleave=4 -mtriple=x86_64-unknown-linux -S -pass-remarks='loop-vectorize' 2>&1 | FileCheck -check-prefix=UNROLLED %s 3 ; RUN: opt < %s -loop-vectorize -force-vector-width=1 -force-vector-interleave=1 -mtriple=x86_64-unknown-linux -S -pass-remarks-analysis='loop-vectorize' 2>&1 | FileCheck -check-prefix=NONE %s 4 5 ; RUN: llc < %s -mtriple x86_64-pc-linux-gnu -o - | FileCheck -check-prefix=DEBUG-OUTPUT %s 6 ; DEBUG-OUTPUT-NOT: .loc 7 ; DEBUG-OUTPUT-NOT: {{.*}}.debug_info 8 9 ; VECTORIZED: remark: vectorization-remarks.c:17:8: vectorized loop (vectorization width: 4, interleaved count: 1) 10 ; UNROLLED: remark: vectorization-remarks.c:17:8: interleaved loop (interleaved count: 4) 11 ; NONE: remark: vectorization-remarks.c:17:8: loop not vectorized: vectorization and interleaving are explicitly disabled, or vectorize width and interleave count are both set to 1 12 13 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" 14 15 define i32 @foo(i32 %n) #0 !dbg !4 { 16 entry: 17 %diff = alloca i32, align 4 18 %cb = alloca [16 x i8], align 16 19 %cc = alloca [16 x i8], align 16 20 store i32 0, i32* %diff, align 4, !tbaa !11 21 br label %for.body 22 23 for.body: ; preds = %for.body, %entry 24 %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] 25 %add8 = phi i32 [ 0, %entry ], [ %add, %for.body ] 26 %arrayidx = getelementptr inbounds [16 x i8], [16 x i8]* %cb, i64 0, i64 %indvars.iv 27 %0 = load i8, i8* %arrayidx, align 1, !tbaa !21 28 %conv = sext i8 %0 to i32 29 %arrayidx2 = getelementptr inbounds [16 x i8], [16 x i8]* %cc, i64 0, i64 %indvars.iv 30 %1 = load i8, i8* %arrayidx2, align 1, !tbaa !21 31 %conv3 = sext i8 %1 to i32 32 %sub = sub i32 %conv, %conv3 33 %add = add nsw i32 %sub, %add8 34 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1 35 %exitcond = icmp eq i64 %indvars.iv.next, 16 36 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !25 37 38 for.end: ; preds = %for.body 39 store i32 %add, i32* %diff, align 4, !tbaa !11 40 call void @ibar(i32* %diff) #2 41 ret i32 0 42 } 43 44 declare void @ibar(i32*) #1 45 46 !llvm.module.flags = !{!7, !8} 47 !llvm.ident = !{!9} 48 !llvm.dbg.cu = !{!24} 49 50 !1 = !DIFile(filename: "vectorization-remarks.c", directory: ".") 51 !2 = !{} 52 !3 = !{!4} 53 !4 = distinct !DISubprogram(name: "foo", line: 5, isLocal: false, isDefinition: true, virtualIndex: 6, flags: DIFlagPrototyped, isOptimized: true, unit: !24, scopeLine: 6, file: !1, scope: !5, type: !6, variables: !2) 54 !5 = !DIFile(filename: "vectorization-remarks.c", directory: ".") 55 !6 = !DISubroutineType(types: !2) 56 !7 = !{i32 2, !"Dwarf Version", i32 4} 57 !8 = !{i32 1, !"Debug Info Version", i32 3} 58 !9 = !{!"clang version 3.5.0 "} 59 !10 = !DILocation(line: 8, column: 3, scope: !4) 60 !11 = !{!12, !12, i64 0} 61 !12 = !{!"int", !13, i64 0} 62 !13 = !{!"omnipotent char", !14, i64 0} 63 !14 = !{!"Simple C/C++ TBAA"} 64 !15 = !DILocation(line: 17, column: 8, scope: !16) 65 !16 = distinct !DILexicalBlock(line: 17, column: 8, file: !1, scope: !17) 66 !17 = distinct !DILexicalBlock(line: 17, column: 8, file: !1, scope: !18) 67 !18 = distinct !DILexicalBlock(line: 17, column: 3, file: !1, scope: !4) 68 !19 = !DILocation(line: 18, column: 5, scope: !20) 69 !20 = distinct !DILexicalBlock(line: 17, column: 27, file: !1, scope: !18) 70 !21 = !{!13, !13, i64 0} 71 !22 = !DILocation(line: 20, column: 3, scope: !4) 72 !23 = !DILocation(line: 21, column: 3, scope: !4) 73 !24 = distinct !DICompileUnit(language: DW_LANG_C89, file: !1, emissionKind: NoDebug) 74 !25 = !{!25, !15} 75
