reference, declarationdefinition
definition → references, declarations, derived classes, virtual overrides
reference to multiple definitions → definitions
unreferenced
    1
    2
    3
    4
    5
    6
    7
    8
    9
   10
   11
   12
   13
   14
   15
   16
   17
   18
   19
   20
   21
   22
   23
   24
   25
   26
   27
   28
   29
   30
   31
   32
   33
   34
   35
   36
   37
   38
   39
   40
   41
   42
   43
   44
   45
   46
   47
   48
   49
   50
   51
   52
   53
   54
   55
   56
   57
   58
   59
   60
   61
   62
   63
   64
   65
   66
   67
   68
   69
   70
   71
   72
   73
   74
   75
   76
   77
   78
   79
   80
   81
   82
   83
   84
   85
   86
   87
   88
   89
   90
   91
   92
   93
   94
   95
   96
   97
   98
   99
  100
  101
  102
  103
  104
  105
  106
  107
  108
  109
  110
  111
  112
  113
  114
  115
  116
  117
  118
  119
  120
  121
  122
  123
  124
  125
  126
  127
  128
  129
  130
  131
  132
  133
  134
  135
  136
  137
  138
  139
  140
  141
  142
  143
; REQUIRES: asserts
; RUN: opt < %s -loop-vectorize -force-vector-width=2 -force-vector-interleave=1 -instcombine -debug-only=loop-vectorize -disable-output -print-after=instcombine 2>&1 | FileCheck %s

target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"

; CHECK-LABEL: vector_gep
; CHECK-NOT:   LV: Found scalar instruction: %tmp0 = getelementptr inbounds i32, i32* %b, i64 %i
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
; CHECK-NEXT:    [[VEC_IND:%.*]] = phi <2 x i64> [ <i64 0, i64 1>, %vector.ph ], [ [[VEC_IND_NEXT:%.*]], %vector.body ]
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32, i32* %b, <2 x i64> [[VEC_IND]]
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[INDEX]]
; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32** [[TMP2]] to <2 x i32*>*
; CHECK-NEXT:    store <2 x i32*> [[TMP1]], <2 x i32*>* [[TMP3]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i64> [[VEC_IND]], <i64 2, i64 2>
; CHECK:         br i1 {{.*}}, label %middle.block, label %vector.body
;
define void @vector_gep(i32** %a, i32 *%b, i64 %n) {
entry:
  br label %for.body

for.body:
  %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
  %tmp0 = getelementptr inbounds i32, i32* %b, i64 %i
  %tmp1 = getelementptr inbounds i32*, i32** %a, i64 %i
  store i32* %tmp0, i32** %tmp1, align 8
  %i.next = add nuw nsw i64 %i, 1
  %cond = icmp slt i64 %i.next, %n
  br i1 %cond, label %for.body, label %for.end

for.end:
  ret void
}

; CHECK-LABEL: scalar_store
; CHECK:       LV: Found scalar instruction: %tmp1 = getelementptr inbounds i32*, i32** %a, i64 %i
; CHECK-NEXT:  LV: Found scalar instruction: %tmp0 = getelementptr inbounds i32, i32* %b, i64 %i
; CHECK-NEXT:  LV: Found scalar instruction: %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
; CHECK-NEXT:  LV: Found scalar instruction: %i.next = add nuw nsw i64 %i, 2
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = shl i64 [[INDEX]], 1
; CHECK-NEXT:    [[TMP4:%.*]] = or i64 [[OFFSET_IDX]], 2
; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* %b, i64 [[OFFSET_IDX]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* %b, i64 [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[OFFSET_IDX]]
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[TMP4]]
; CHECK-NEXT:    store i32* [[TMP5]], i32** [[TMP7]], align 8
; CHECK-NEXT:    store i32* [[TMP6]], i32** [[TMP8]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; CHECK:         br i1 {{.*}}, label %middle.block, label %vector.body
;
define void @scalar_store(i32** %a, i32 *%b, i64 %n) {
entry:
  br label %for.body

for.body:
  %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
  %tmp0 = getelementptr inbounds i32, i32* %b, i64 %i
  %tmp1 = getelementptr inbounds i32*, i32** %a, i64 %i
  store i32* %tmp0, i32** %tmp1, align 8
  %i.next = add nuw nsw i64 %i, 2
  %cond = icmp slt i64 %i.next, %n
  br i1 %cond, label %for.body, label %for.end

for.end:
  ret void
}

; CHECK-LABEL: expansion
; CHECK:       LV: Found scalar instruction: %tmp3 = getelementptr inbounds i32*, i32** %tmp2, i64 %i
; CHECK-NEXT:  LV: Found scalar instruction: %tmp1 = bitcast i64* %tmp0 to i32*
; CHECK-NEXT:  LV: Found scalar instruction: %tmp2 = getelementptr inbounds i32*, i32** %a, i64 0
; CHECK-NEXT:  LV: Found scalar instruction: %tmp0 = getelementptr inbounds i64, i64* %b, i64 %i
; CHECK-NEXT:  LV: Found scalar instruction: %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
; CHECK-NEXT:  LV: Found scalar instruction: %i.next = add nuw nsw i64 %i, 2
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = shl i64 [[INDEX]], 1
; CHECK-NEXT:    [[TMP4:%.*]] = or i64 [[OFFSET_IDX]], 2
; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i64, i64* %b, i64 [[OFFSET_IDX]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i64, i64* %b, i64 [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[OFFSET_IDX]]
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[TMP4]]
; CHECK-NEXT:    [[TMP9:%.*]] = bitcast i32** [[TMP7]] to i64**
; CHECK-NEXT:    store i64* [[TMP5]], i64** [[TMP9]], align 8
; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i32** [[TMP8]] to i64**
; CHECK-NEXT:    store i64* [[TMP6]], i64** [[TMP10]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; CHECK:         br i1 {{.*}}, label %middle.block, label %vector.body
;
define void @expansion(i32** %a, i64 *%b, i64 %n) {
entry:
  br label %for.body

for.body:
  %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
  %tmp0 = getelementptr inbounds i64, i64* %b, i64 %i
  %tmp1 = bitcast i64* %tmp0 to i32*
  %tmp2 = getelementptr inbounds i32*, i32** %a, i64 0
  %tmp3 = getelementptr inbounds i32*, i32** %tmp2, i64 %i
  store i32* %tmp1, i32** %tmp3, align 8
  %i.next = add nuw nsw i64 %i, 2
  %cond = icmp slt i64 %i.next, %n
  br i1 %cond, label %for.body, label %for.end

for.end:
  ret void
}

; CHECK-LABEL: no_gep_or_bitcast
; CHECK-NOT:   LV: Found scalar instruction: %tmp1 = load i32*, i32** %tmp0, align 8
; CHECK:       LV: Found scalar instruction: %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
; CHECK-NEXT:  LV: Found scalar instruction: %i.next = add nuw nsw i64 %i, 1
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32*, i32** %a, i64 [[INDEX]]
; CHECK-NEXT:    [[TMP2:%.*]] = bitcast i32** [[TMP1]] to <2 x i32*>*
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <2 x i32*>, <2 x i32*>* [[TMP2]], align 8
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x i32*> [[WIDE_LOAD]], i32 0
; CHECK-NEXT:    store i32 0, i32* [[TMP3]], align 8
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x i32*> [[WIDE_LOAD]], i32 1
; CHECK-NEXT:    store i32 0, i32* [[TMP4]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
; CHECK:         br i1 {{.*}}, label %middle.block, label %vector.body
;
define void @no_gep_or_bitcast(i32** noalias %a, i64 %n) {
entry:
  br label %for.body

for.body:
  %i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
  %tmp0 = getelementptr inbounds i32*, i32** %a, i64 %i
  %tmp1 = load i32*, i32** %tmp0, align 8
  store i32 0, i32* %tmp1, align 8
  %i.next = add nuw nsw i64 %i, 1
  %cond = icmp slt i64 %i.next, %n
  br i1 %cond, label %for.body, label %for.end

for.end:
  ret void
}