1 ; RUN: opt -loop-distribute -S < %s 2>&1 \ 2 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=NO_REMARKS 3 ; RUN: opt -loop-distribute -S -pass-remarks-missed=loop-distribute < %s 2>&1 \ 4 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=MISSED_REMARKS 5 ; RUN: opt -loop-distribute -S -pass-remarks-analysis=loop-distribute < %s 2>&1 \ 6 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=ANALYSIS_REMARKS 7 ; RUN: opt -loop-distribute -S -pass-remarks=loop-distribute < %s 2>&1 \ 8 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=REMARKS 9 10 ; This is the input program: 11 ; 12 ; 1 void forced (char *A, char *B, char *C, int N) { 13 ; 2 #pragma clang loop distribute(enable) 14 ; 3 for(int i = 0; i < N; i++) { 15 ; 4 A[i] = B[i] * C[i]; 16 ; 5 } 17 ; 6 } 18 ; 7 19 ; 8 void not_forced (char *A, char *B, char *C, int N) { 20 ; 9 for(int i = 0; i < N; i++) { 21 ; 10 A[i] = B[i] * C[i]; 22 ; 11 } 23 ; 12 } 24 ; 13 25 ; 14 void success (char *A, char *B, char *C, char *D, char *E, int N) { 26 ; 15 for(int i = 0; i < N; i++) { 27 ; 16 A[i + 1] = A[i] + B[i]; 28 ; 17 C[i] = D[i] * E[i]; 29 ; 18 } 30 ; 19 } 31 32 target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" 33 target triple = "x86_64-apple-macosx10.11.0" 34 35 ; MISSED_REMARKS: remark: /tmp/t.c:3:3: loop not distributed: use -Rpass-analysis=loop-distribute for more info 36 ; ALWAYS: remark: /tmp/t.c:3:3: loop not distributed: memory operations are safe for vectorization 37 ; ALWAYS: warning: /tmp/t.c:3:3: loop not disributed: failed explicitly specified loop distribution 38 39 define void @forced(i8* %A, i8* %B, i8* %C, i32 %N) !dbg !7 { 40 entry: 41 %cmp12 = icmp sgt i32 %N, 0, !dbg !9 42 br i1 %cmp12, label %ph, label %for.cond.cleanup, !dbg !10 43 44 ph: 45 br label %for.body 46 47 for.body: 48 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 49 %arrayidx = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !12 50 %0 = load i8, i8* %arrayidx, align 1, !dbg !12, !tbaa !13 51 %arrayidx2 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !16 52 %1 = load i8, i8* %arrayidx2, align 1, !dbg !16, !tbaa !13 53 %mul = mul i8 %1, %0, !dbg !17 54 %arrayidx6 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !18 55 store i8 %mul, i8* %arrayidx6, align 1, !dbg !19, !tbaa !13 56 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !10 57 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !10 58 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !10 59 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !10, !llvm.loop !20 60 61 for.cond.cleanup: 62 ret void, !dbg !11 63 } 64 65 ; NO_REMARKS-NOT: remark: /tmp/t.c:9:3: loop not distributed: memory operations are safe for vectorization 66 ; MISSED_REMARKS: remark: /tmp/t.c:9:3: loop not distributed: use -Rpass-analysis=loop-distribute for more info 67 ; ANALYSIS_REMARKS: remark: /tmp/t.c:9:3: loop not distributed: memory operations are safe for vectorization 68 ; ALWAYS-NOT: warning: /tmp/t.c:9:3: loop not disributed: failed explicitly specified loop distribution 69 70 define void @not_forced(i8* %A, i8* %B, i8* %C, i32 %N) !dbg !22 { 71 entry: 72 %cmp12 = icmp sgt i32 %N, 0, !dbg !23 73 br i1 %cmp12, label %ph, label %for.cond.cleanup, !dbg !24 74 75 ph: 76 br label %for.body 77 78 for.body: 79 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 80 %arrayidx = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !26 81 %0 = load i8, i8* %arrayidx, align 1, !dbg !26, !tbaa !13 82 %arrayidx2 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !27 83 %1 = load i8, i8* %arrayidx2, align 1, !dbg !27, !tbaa !13 84 %mul = mul i8 %1, %0, !dbg !28 85 %arrayidx6 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !29 86 store i8 %mul, i8* %arrayidx6, align 1, !dbg !30, !tbaa !13 87 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !24 88 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !24 89 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !24 90 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !24 91 92 for.cond.cleanup: 93 ret void, !dbg !25 94 } 95 96 ; REMARKS: remark: /tmp/t.c:15:3: distributed loop 97 98 define void @success(i8* %A, i8* %B, i8* %C, i8* %D, i8* %E, i32 %N) !dbg !31 { 99 entry: 100 %cmp28 = icmp sgt i32 %N, 0, !dbg !32 101 br i1 %cmp28, label %ph, label %for.cond.cleanup, !dbg !33 102 103 ph: 104 br label %for.body 105 106 for.body: 107 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 108 %arrayidx = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !35 109 %0 = load i8, i8* %arrayidx, align 1, !dbg !35, !tbaa !13 110 %arrayidx2 = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !36 111 %1 = load i8, i8* %arrayidx2, align 1, !dbg !36, !tbaa !13 112 %add = add i8 %1, %0, !dbg !37 113 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !33 114 %arrayidx7 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv.next, !dbg !38 115 store i8 %add, i8* %arrayidx7, align 1, !dbg !39, !tbaa !13 116 %arrayidx9 = getelementptr inbounds i8, i8* %D, i64 %indvars.iv, !dbg !40 117 %2 = load i8, i8* %arrayidx9, align 1, !dbg !40, !tbaa !13 118 %arrayidx12 = getelementptr inbounds i8, i8* %E, i64 %indvars.iv, !dbg !41 119 %3 = load i8, i8* %arrayidx12, align 1, !dbg !41, !tbaa !13 120 %mul = mul i8 %3, %2, !dbg !42 121 %arrayidx16 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !43 122 store i8 %mul, i8* %arrayidx16, align 1, !dbg !44, !tbaa !13 123 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !33 124 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !33 125 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !33 126 127 for.cond.cleanup: 128 ret void, !dbg !34 129 } 130 131 !llvm.dbg.cu = !{!0} 132 !llvm.module.flags = !{!3, !4} 133 134 !0 = distinct !DICompileUnit(language: DW_LANG_C99, file: !1, producer: "clang version 3.9.0 (trunk 267633) (llvm/trunk 267675)", isOptimized: true, runtimeVersion: 0, emissionKind: LineTablesOnly, enums: !2) 135 !1 = !DIFile(filename: "/tmp/t.c", directory: "/tmp") 136 !2 = !{} 137 !3 = !{i32 2, !"Dwarf Version", i32 2} 138 !4 = !{i32 2, !"Debug Info Version", i32 3} 139 !7 = distinct !DISubprogram(name: "forced", scope: !1, file: !1, line: 1, type: !8, isLocal: false, isDefinition: true, scopeLine: 1, flags: DIFlagPrototyped, isOptimized: true, unit: !0, variables: !2) 140 !8 = !DISubroutineType(types: !2) 141 !9 = !DILocation(line: 3, column: 20, scope: !7) 142 !10 = !DILocation(line: 3, column: 3, scope: !7) 143 !11 = !DILocation(line: 6, column: 1, scope: !7) 144 !12 = !DILocation(line: 4, column: 12, scope: !7) 145 !13 = !{!14, !14, i64 0} 146 !14 = !{!"omnipotent char", !15, i64 0} 147 !15 = !{!"Simple C/C++ TBAA"} 148 !16 = !DILocation(line: 4, column: 19, scope: !7) 149 !17 = !DILocation(line: 4, column: 17, scope: !7) 150 !18 = !DILocation(line: 4, column: 5, scope: !7) 151 !19 = !DILocation(line: 4, column: 10, scope: !7) 152 !20 = distinct !{!20, !21} 153 !21 = !{!"llvm.loop.distribute.enable", i1 true} 154 !22 = distinct !DISubprogram(name: "not_forced", scope: !1, file: !1, line: 8, type: !8, isLocal: false, isDefinition: true, scopeLine: 8, flags: DIFlagPrototyped, isOptimized: true, unit: !0, variables: !2) 155 !23 = !DILocation(line: 9, column: 20, scope: !22) 156 !24 = !DILocation(line: 9, column: 3, scope: !22) 157 !25 = !DILocation(line: 12, column: 1, scope: !22) 158 !26 = !DILocation(line: 10, column: 12, scope: !22) 159 !27 = !DILocation(line: 10, column: 19, scope: !22) 160 !28 = !DILocation(line: 10, column: 17, scope: !22) 161 !29 = !DILocation(line: 10, column: 5, scope: !22) 162 !30 = !DILocation(line: 10, column: 10, scope: !22) 163 !31 = distinct !DISubprogram(name: "success", scope: !1, file: !1, line: 14, type: !8, isLocal: false, isDefinition: true, scopeLine: 14, flags: DIFlagPrototyped, isOptimized: true, unit: !0, variables: !2) 164 !32 = !DILocation(line: 15, column: 20, scope: !31) 165 !33 = !DILocation(line: 15, column: 3, scope: !31) 166 !34 = !DILocation(line: 19, column: 1, scope: !31) 167 !35 = !DILocation(line: 16, column: 16, scope: !31) 168 !36 = !DILocation(line: 16, column: 23, scope: !31) 169 !37 = !DILocation(line: 16, column: 21, scope: !31) 170 !38 = !DILocation(line: 16, column: 5, scope: !31) 171 !39 = !DILocation(line: 16, column: 14, scope: !31) 172 !40 = !DILocation(line: 17, column: 12, scope: !31) 173 !41 = !DILocation(line: 17, column: 19, scope: !31) 174 !42 = !DILocation(line: 17, column: 17, scope: !31) 175 !43 = !DILocation(line: 17, column: 5, scope: !31) 176 !44 = !DILocation(line: 17, column: 10, scope: !31) 177
