Mercurial > hg > CbC > CbC_llvm

comparison test/Transforms/LoopVectorize/runtime-check.ll @ 95:afa8332a0e37 LLVM3.8

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
LLVM 3.8
author Kaito Tokumori <e105711@ie.u-ryukyu.ac.jp>
date Tue, 13 Oct 2015 17:48:58 +0900
parents 60c9769439b8
children 1172e4bd9c6f
comparison
equal deleted inserted replaced
84:f3e34b893a5f 95:afa8332a0e37
7 ; int foo(float *a, float *b, int n) { 7 ; int foo(float *a, float *b, int n) {
8 ; for (int i=0; i<n; ++i) 8 ; for (int i=0; i<n; ++i)
9 ; a[i] = b[i] * 3; 9 ; a[i] = b[i] * 3;
10 ; } 10 ; }
11 11
12 ;CHECK-LABEL: define i32 @foo
12 ;CHECK: for.body.preheader: 13 ;CHECK: for.body.preheader:
13 ;CHECK: br i1 %cmp.zero, label %middle.block, label %vector.memcheck 14 ;CHECK: br i1 %cmp.zero, label %scalar.ph, label %vector.memcheck, !dbg [[BODY_LOC:![0-9]+]]
14 ;CHECK: vector.memcheck: 15 ;CHECK: vector.memcheck:
15 ;CHECK: br i1 %memcheck.conflict, label %middle.block, label %vector.ph 16 ;CHECK: br i1 %memcheck.conflict, label %scalar.ph, label %vector.ph, !dbg [[BODY_LOC]]
16 ;CHECK: load <4 x float> 17 ;CHECK: load <4 x float>
17 define i32 @foo(float* nocapture %a, float* nocapture %b, i32 %n) nounwind uwtable ssp { 18 define i32 @foo(float* nocapture %a, float* nocapture %b, i32 %n) nounwind uwtable ssp {
18 entry: 19 entry:
19 %cmp6 = icmp sgt i32 %n, 0 20 %cmp6 = icmp sgt i32 %n, 0, !dbg !6
20 br i1 %cmp6, label %for.body, label %for.end 21 br i1 %cmp6, label %for.body, label %for.end, !dbg !6
21 22
22 for.body: ; preds = %entry, %for.body 23 for.body: ; preds = %entry, %for.body
23 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ] 24 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ], !dbg !7
24 %arrayidx = getelementptr inbounds float* %b, i64 %indvars.iv 25 %arrayidx = getelementptr inbounds float, float* %b, i64 %indvars.iv, !dbg !7
25 %0 = load float* %arrayidx, align 4 26 %0 = load float, float* %arrayidx, align 4, !dbg !7
26 %mul = fmul float %0, 3.000000e+00 27 %mul = fmul float %0, 3.000000e+00, !dbg !7
27 %arrayidx2 = getelementptr inbounds float* %a, i64 %indvars.iv 28 %arrayidx2 = getelementptr inbounds float, float* %a, i64 %indvars.iv, !dbg !7
28 store float %mul, float* %arrayidx2, align 4 29 store float %mul, float* %arrayidx2, align 4, !dbg !7
29 %indvars.iv.next = add i64 %indvars.iv, 1 30 %indvars.iv.next = add i64 %indvars.iv, 1, !dbg !7
30 %lftr.wideiv = trunc i64 %indvars.iv.next to i32 31 %lftr.wideiv = trunc i64 %indvars.iv.next to i32, !dbg !7
31 %exitcond = icmp eq i32 %lftr.wideiv, %n 32 %exitcond = icmp eq i32 %lftr.wideiv, %n, !dbg !7
32 br i1 %exitcond, label %for.end, label %for.body 33 br i1 %exitcond, label %for.end, label %for.body, !dbg !7
33 34
34 for.end: ; preds = %for.body, %entry 35 for.end: ; preds = %for.body, %entry
35 ret i32 undef 36 ret i32 undef, !dbg !8
36 } 37 }
37 38
38 ; Make sure that we try to vectorize loops with a runtime check if the 39 ; Make sure that we try to vectorize loops with a runtime check if the
39 ; dependency check fails. 40 ; dependency check fails.
40 41
45 br label %for.body 46 br label %for.body
46 47
47 for.body: 48 for.body:
48 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] 49 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
49 %ind.sum = add i64 %iv, %offset 50 %ind.sum = add i64 %iv, %offset
50 %arr.idx = getelementptr inbounds float* %a, i64 %ind.sum 51 %arr.idx = getelementptr inbounds float, float* %a, i64 %ind.sum
51 %l1 = load float* %arr.idx, align 4 52 %l1 = load float, float* %arr.idx, align 4
52 %ind.sum2 = add i64 %iv, %offset2 53 %ind.sum2 = add i64 %iv, %offset2
53 %arr.idx2 = getelementptr inbounds float* %a, i64 %ind.sum2 54 %arr.idx2 = getelementptr inbounds float, float* %a, i64 %ind.sum2
54 %l2 = load float* %arr.idx2, align 4 55 %l2 = load float, float* %arr.idx2, align 4
55 %m = fmul fast float %b, %l2 56 %m = fmul fast float %b, %l2
56 %ad = fadd fast float %l1, %m 57 %ad = fadd fast float %l1, %m
57 store float %ad, float* %arr.idx, align 4 58 store float %ad, float* %arr.idx, align 4
58 %iv.next = add nuw nsw i64 %iv, 1 59 %iv.next = add nuw nsw i64 %iv, 1
59 %exitcond = icmp eq i64 %iv.next, %n 60 %exitcond = icmp eq i64 %iv.next, %n
60 br i1 %exitcond, label %loopexit, label %for.body 61 br i1 %exitcond, label %loopexit, label %for.body
61 62
62 loopexit: 63 loopexit:
63 ret void 64 ret void
64 } 65 }
66
67 ; CHECK: [[BODY_LOC]] = !DILocation(line: 101, column: 1, scope: !{{.*}})
68
69 !llvm.module.flags = !{!0, !1}
70 !0 = !{i32 2, !"Dwarf Version", i32 4}
71 !1 = !{i32 2, !"Debug Info Version", i32 3}
72
73 !2 = !{}
74 !3 = !DISubroutineType(types: !2)
75 !4 = !DIFile(filename: "test.cpp", directory: "/tmp")
76 !5 = distinct !DISubprogram(name: "foo", scope: !4, file: !4, line: 99, type: !3, isLocal: false, isDefinition: true, scopeLine: 100, flags: DIFlagPrototyped, isOptimized: false, variables: !2)
77 !6 = !DILocation(line: 100, column: 1, scope: !5)
78 !7 = !DILocation(line: 101, column: 1, scope: !5)
79 !8 = !DILocation(line: 102, column: 1, scope: !5)