1 ; RUN: opt -loop-simplify -loop-distribute -enable-loop-distribute -S < %s 2>&1 \ 2 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=NO_REMARKS 3 ; RUN: opt -loop-simplify -loop-distribute -enable-loop-distribute -S \ 4 ; RUN: -pass-remarks-missed=loop-distribute < %s 2>&1 \ 5 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=MISSED_REMARKS 6 ; RUN: opt -loop-simplify -loop-distribute -enable-loop-distribute -S \ 7 ; RUN: -pass-remarks-analysis=loop-distribute < %s 2>&1 \ 8 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=ANALYSIS_REMARKS 9 ; RUN: opt -loop-simplify -loop-distribute -enable-loop-distribute -S \ 10 ; RUN: -pass-remarks=loop-distribute < %s 2>&1 \ 11 ; RUN: | FileCheck %s --check-prefix=ALWAYS --check-prefix=REMARKS 12 13 ; This is the input program: 14 ; 15 ; 1 void forced (char *A, char *B, char *C, int N) { 16 ; 2 #pragma clang loop distribute(enable) 17 ; 3 for(int i = 0; i < N; i++) { 18 ; 4 A[i] = B[i] * C[i]; 19 ; 5 } 20 ; 6 } 21 ; 7 22 ; 8 void not_forced (char *A, char *B, char *C, int N) { 23 ; 9 for(int i = 0; i < N; i++) { 24 ; 10 A[i] = B[i] * C[i]; 25 ; 11 } 26 ; 12 } 27 ; 13 28 ; 14 void success (char *A, char *B, char *C, char *D, char *E, int N) { 29 ; 15 for(int i = 0; i < N; i++) { 30 ; 16 A[i + 1] = A[i] + B[i]; 31 ; 17 C[i] = D[i] * E[i]; 32 ; 18 } 33 ; 19 } 34 35 target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" 36 target triple = "x86_64-apple-macosx10.11.0" 37 38 ; MISSED_REMARKS: remark: /tmp/t.c:3:3: loop not distributed: use -Rpass-analysis=loop-distribute for more info 39 ; ALWAYS: remark: /tmp/t.c:3:3: loop not distributed: memory operations are safe for vectorization 40 ; ALWAYS: warning: /tmp/t.c:3:3: loop not distributed: failed explicitly specified loop distribution 41 42 define void @forced(i8* %A, i8* %B, i8* %C, i32 %N) !dbg !7 { 43 entry: 44 %cmp12 = icmp sgt i32 %N, 0, !dbg !9 45 br i1 %cmp12, label %ph, label %for.cond.cleanup, !dbg !10 46 47 ph: 48 br label %for.body 49 50 for.body: 51 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 52 %arrayidx = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !12 53 %0 = load i8, i8* %arrayidx, align 1, !dbg !12, !tbaa !13 54 %arrayidx2 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !16 55 %1 = load i8, i8* %arrayidx2, align 1, !dbg !16, !tbaa !13 56 %mul = mul i8 %1, %0, !dbg !17 57 %arrayidx6 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !18 58 store i8 %mul, i8* %arrayidx6, align 1, !dbg !19, !tbaa !13 59 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !10 60 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !10 61 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !10 62 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !10, !llvm.loop !20 63 64 for.cond.cleanup: 65 ret void, !dbg !11 66 } 67 68 ; NO_REMARKS-NOT: remark: /tmp/t.c:9:3: loop not distributed: memory operations are safe for vectorization 69 ; MISSED_REMARKS: remark: /tmp/t.c:9:3: loop not distributed: use -Rpass-analysis=loop-distribute for more info 70 ; ANALYSIS_REMARKS: remark: /tmp/t.c:9:3: loop not distributed: memory operations are safe for vectorization 71 ; ALWAYS-NOT: warning: /tmp/t.c:9:3: loop not distributed: failed explicitly specified loop distribution 72 73 define void @not_forced(i8* %A, i8* %B, i8* %C, i32 %N) !dbg !22 { 74 entry: 75 %cmp12 = icmp sgt i32 %N, 0, !dbg !23 76 br i1 %cmp12, label %ph, label %for.cond.cleanup, !dbg !24 77 78 ph: 79 br label %for.body 80 81 for.body: 82 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 83 %arrayidx = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !26 84 %0 = load i8, i8* %arrayidx, align 1, !dbg !26, !tbaa !13 85 %arrayidx2 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !27 86 %1 = load i8, i8* %arrayidx2, align 1, !dbg !27, !tbaa !13 87 %mul = mul i8 %1, %0, !dbg !28 88 %arrayidx6 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !29 89 store i8 %mul, i8* %arrayidx6, align 1, !dbg !30, !tbaa !13 90 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !24 91 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !24 92 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !24 93 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !24 94 95 for.cond.cleanup: 96 ret void, !dbg !25 97 } 98 99 ; REMARKS: remark: /tmp/t.c:15:3: distributed loop 100 101 define void @success(i8* %A, i8* %B, i8* %C, i8* %D, i8* %E, i32 %N) !dbg !31 { 102 entry: 103 %cmp28 = icmp sgt i32 %N, 0, !dbg !32 104 br i1 %cmp28, label %ph, label %for.cond.cleanup, !dbg !33 105 106 ph: 107 br label %for.body 108 109 for.body: 110 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %ph ] 111 %arrayidx = getelementptr inbounds i8, i8* %A, i64 %indvars.iv, !dbg !35 112 %0 = load i8, i8* %arrayidx, align 1, !dbg !35, !tbaa !13 113 %arrayidx2 = getelementptr inbounds i8, i8* %B, i64 %indvars.iv, !dbg !36 114 %1 = load i8, i8* %arrayidx2, align 1, !dbg !36, !tbaa !13 115 %add = add i8 %1, %0, !dbg !37 116 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1, !dbg !33 117 %arrayidx7 = getelementptr inbounds i8, i8* %A, i64 %indvars.iv.next, !dbg !38 118 store i8 %add, i8* %arrayidx7, align 1, !dbg !39, !tbaa !13 119 %arrayidx9 = getelementptr inbounds i8, i8* %D, i64 %indvars.iv, !dbg !40 120 %2 = load i8, i8* %arrayidx9, align 1, !dbg !40, !tbaa !13 121 %arrayidx12 = getelementptr inbounds i8, i8* %E, i64 %indvars.iv, !dbg !41 122 %3 = load i8, i8* %arrayidx12, align 1, !dbg !41, !tbaa !13 123 %mul = mul i8 %3, %2, !dbg !42 124 %arrayidx16 = getelementptr inbounds i8, i8* %C, i64 %indvars.iv, !dbg !43 125 store i8 %mul, i8* %arrayidx16, align 1, !dbg !44, !tbaa !13 126 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !33 127 %exitcond = icmp eq i32 %lftr.wideiv, %N, !dbg !33 128 br i1 %exitcond, label %for.cond.cleanup, label %for.body, !dbg !33 129 130 for.cond.cleanup: 131 ret void, !dbg !34 132 } 133 134 !llvm.dbg.cu = !{!0} 135 !llvm.module.flags = !{!3, !4} 136 137 !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) 138 !1 = !DIFile(filename: "/tmp/t.c", directory: "/tmp") 139 !2 = !{} 140 !3 = !{i32 2, !"Dwarf Version", i32 2} 141 !4 = !{i32 2, !"Debug Info Version", i32 3} 142 !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, retainedNodes: !2) 143 !8 = !DISubroutineType(types: !2) 144 !9 = !DILocation(line: 3, column: 20, scope: !7) 145 !10 = !DILocation(line: 3, column: 3, scope: !7) 146 !11 = !DILocation(line: 6, column: 1, scope: !7) 147 !12 = !DILocation(line: 4, column: 12, scope: !7) 148 !13 = !{!14, !14, i64 0} 149 !14 = !{!"omnipotent char", !15, i64 0} 150 !15 = !{!"Simple C/C++ TBAA"} 151 !16 = !DILocation(line: 4, column: 19, scope: !7) 152 !17 = !DILocation(line: 4, column: 17, scope: !7) 153 !18 = !DILocation(line: 4, column: 5, scope: !7) 154 !19 = !DILocation(line: 4, column: 10, scope: !7) 155 !20 = distinct !{!20, !21} 156 !21 = !{!"llvm.loop.distribute.enable", i1 true} 157 !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, retainedNodes: !2) 158 !23 = !DILocation(line: 9, column: 20, scope: !22) 159 !24 = !DILocation(line: 9, column: 3, scope: !22) 160 !25 = !DILocation(line: 12, column: 1, scope: !22) 161 !26 = !DILocation(line: 10, column: 12, scope: !22) 162 !27 = !DILocation(line: 10, column: 19, scope: !22) 163 !28 = !DILocation(line: 10, column: 17, scope: !22) 164 !29 = !DILocation(line: 10, column: 5, scope: !22) 165 !30 = !DILocation(line: 10, column: 10, scope: !22) 166 !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, retainedNodes: !2) 167 !32 = !DILocation(line: 15, column: 20, scope: !31) 168 !33 = !DILocation(line: 15, column: 3, scope: !31) 169 !34 = !DILocation(line: 19, column: 1, scope: !31) 170 !35 = !DILocation(line: 16, column: 16, scope: !31) 171 !36 = !DILocation(line: 16, column: 23, scope: !31) 172 !37 = !DILocation(line: 16, column: 21, scope: !31) 173 !38 = !DILocation(line: 16, column: 5, scope: !31) 174 !39 = !DILocation(line: 16, column: 14, scope: !31) 175 !40 = !DILocation(line: 17, column: 12, scope: !31) 176 !41 = !DILocation(line: 17, column: 19, scope: !31) 177 !42 = !DILocation(line: 17, column: 17, scope: !31) 178 !43 = !DILocation(line: 17, column: 5, scope: !31) 179 !44 = !DILocation(line: 17, column: 10, scope: !31) 180
