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
144
145
146
147
148
149
150
151
152
153
154
155
156
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 5
; RUN: opt -p loop-vectorize -force-vector-width=2 -force-vector-interleave=2 -S %s | FileCheck %s
target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-n32:64-S128-Fn32"
target triple = "arm64-apple-macosx"
define void @load_store_interleave_group(ptr noalias %data) {
; CHECK-LABEL: define void @load_store_interleave_group(
; CHECK-SAME: ptr noalias [[DATA:%.*]]) {
; CHECK-NEXT: [[ENTRY:.*]]:
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 1
; CHECK-NEXT: [[TMP2:%.*]] = shl nsw i64 [[INDEX]], 1
; CHECK-NEXT: [[TMP3:%.*]] = shl nsw i64 [[TMP1]], 1
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP2]]
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP3]]
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <2 x i64>, ptr [[TMP4]], align 8
; CHECK-NEXT: [[WIDE_LOAD1:%.*]] = load <2 x i64>, ptr [[TMP5]], align 8
; CHECK-NEXT: store <2 x i64> [[WIDE_LOAD]], ptr [[TMP4]], align 8
; CHECK-NEXT: store <2 x i64> [[WIDE_LOAD1]], ptr [[TMP5]], align 8
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT: br i1 [[TMP6]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br i1 true, label %[[EXIT:.*]], label %[[SCALAR_PH]]
; CHECK: [[SCALAR_PH]]:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ 100, %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ]
; CHECK-NEXT: br label %[[LOOP:.*]]
; CHECK: [[LOOP]]:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT: [[MUL_2:%.*]] = shl nsw i64 [[IV]], 1
; CHECK-NEXT: [[DATA_0:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[MUL_2]]
; CHECK-NEXT: [[L_0:%.*]] = load i64, ptr [[DATA_0]], align 8
; CHECK-NEXT: store i64 [[L_0]], ptr [[DATA_0]], align 8
; CHECK-NEXT: [[ADD_1:%.*]] = or disjoint i64 [[MUL_2]], 1
; CHECK-NEXT: [[DATA_1:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[ADD_1]]
; CHECK-NEXT: [[L_1:%.*]] = load i64, ptr [[DATA_1]], align 8
; CHECK-NEXT: store i64 [[L_1]], ptr [[DATA_1]], align 8
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT: [[EC:%.*]] = icmp eq i64 [[IV_NEXT]], 100
; CHECK-NEXT: br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: [[EXIT]]:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%mul.2 = shl nsw i64 %iv, 1
%data.0 = getelementptr inbounds i64, ptr %data, i64 %mul.2
%l.0 = load i64, ptr %data.0, align 8
store i64 %l.0, ptr %data.0, align 8
%add.1 = or disjoint i64 %mul.2, 1
%data.1 = getelementptr inbounds i64, ptr %data, i64 %add.1
%l.1 = load i64, ptr %data.1, align 8
store i64 %l.1, ptr %data.1, align 8
%iv.next = add nuw nsw i64 %iv, 1
%ec = icmp eq i64 %iv.next, 100
br i1 %ec, label %exit, label %loop
exit:
ret void
}
define void @test_2xi64_with_wide_load(ptr noalias %data, ptr noalias %factor) {
; CHECK-LABEL: define void @test_2xi64_with_wide_load(
; CHECK-SAME: ptr noalias [[DATA:%.*]], ptr noalias [[FACTOR:%.*]]) {
; CHECK-NEXT: [[ENTRY:.*]]:
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i64, ptr [[FACTOR]], i64 [[INDEX]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i64, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i64, ptr [[TMP1]], i32 2
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = load <2 x i64>, ptr [[TMP2]], align 8
; CHECK-NEXT: [[BROADCAST_SPLAT3:%.*]] = load <2 x i64>, ptr [[TMP3]], align 8
; CHECK-NEXT: [[TMP6:%.*]] = shl nsw i64 [[INDEX]], 1
; CHECK-NEXT: [[TMP7:%.*]] = shl nsw i64 [[TMP0]], 1
; CHECK-NEXT: [[TMP8:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP6]]
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP7]]
; CHECK-NEXT: [[WIDE_VEC:%.*]] = load <4 x i64>, ptr [[TMP8]], align 8
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = shufflevector <4 x i64> [[WIDE_VEC]], <4 x i64> poison, <2 x i32> <i32 0, i32 2>
; CHECK-NEXT: [[STRIDED_VEC2:%.*]] = shufflevector <4 x i64> [[WIDE_VEC]], <4 x i64> poison, <2 x i32> <i32 1, i32 3>
; CHECK-NEXT: [[WIDE_VEC3:%.*]] = load <4 x i64>, ptr [[TMP9]], align 8
; CHECK-NEXT: [[WIDE_LOAD1:%.*]] = shufflevector <4 x i64> [[WIDE_VEC3]], <4 x i64> poison, <2 x i32> <i32 0, i32 2>
; CHECK-NEXT: [[STRIDED_VEC5:%.*]] = shufflevector <4 x i64> [[WIDE_VEC3]], <4 x i64> poison, <2 x i32> <i32 1, i32 3>
; CHECK-NEXT: [[TMP10:%.*]] = mul <2 x i64> [[BROADCAST_SPLAT]], [[WIDE_LOAD]]
; CHECK-NEXT: [[TMP11:%.*]] = mul <2 x i64> [[BROADCAST_SPLAT3]], [[WIDE_LOAD1]]
; CHECK-NEXT: [[TMP15:%.*]] = mul <2 x i64> [[BROADCAST_SPLAT]], [[STRIDED_VEC2]]
; CHECK-NEXT: [[TMP16:%.*]] = mul <2 x i64> [[BROADCAST_SPLAT3]], [[STRIDED_VEC5]]
; CHECK-NEXT: [[TMP17:%.*]] = shufflevector <2 x i64> [[TMP10]], <2 x i64> [[TMP15]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
; CHECK-NEXT: [[INTERLEAVED_VEC:%.*]] = shufflevector <4 x i64> [[TMP17]], <4 x i64> poison, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
; CHECK-NEXT: store <4 x i64> [[INTERLEAVED_VEC]], ptr [[TMP8]], align 8
; CHECK-NEXT: [[TMP18:%.*]] = shufflevector <2 x i64> [[TMP11]], <2 x i64> [[TMP16]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
; CHECK-NEXT: [[INTERLEAVED_VEC6:%.*]] = shufflevector <4 x i64> [[TMP18]], <4 x i64> poison, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
; CHECK-NEXT: store <4 x i64> [[INTERLEAVED_VEC6]], ptr [[TMP9]], align 8
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT: br i1 [[TMP12]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
; CHECK-NEXT: br i1 true, label %[[EXIT:.*]], label %[[SCALAR_PH]]
; CHECK: [[SCALAR_PH]]:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ 100, %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ]
; CHECK-NEXT: br label %[[LOOP:.*]]
; CHECK: [[LOOP]]:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i64, ptr [[FACTOR]], i64 [[IV]]
; CHECK-NEXT: [[L_FACTOR:%.*]] = load i64, ptr [[ARRAYIDX]], align 8
; CHECK-NEXT: [[TMP13:%.*]] = shl nsw i64 [[IV]], 1
; CHECK-NEXT: [[DATA_0:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP13]]
; CHECK-NEXT: [[L_0:%.*]] = load i64, ptr [[DATA_0]], align 8
; CHECK-NEXT: [[MUL_0:%.*]] = mul i64 [[L_FACTOR]], [[L_0]]
; CHECK-NEXT: store i64 [[MUL_0]], ptr [[DATA_0]], align 8
; CHECK-NEXT: [[TMP14:%.*]] = or disjoint i64 [[TMP13]], 1
; CHECK-NEXT: [[DATA_1:%.*]] = getelementptr inbounds i64, ptr [[DATA]], i64 [[TMP14]]
; CHECK-NEXT: [[L_1:%.*]] = load i64, ptr [[DATA_1]], align 8
; CHECK-NEXT: [[MUL_1:%.*]] = mul i64 [[L_FACTOR]], [[L_1]]
; CHECK-NEXT: store i64 [[MUL_1]], ptr [[DATA_1]], align 8
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT: [[EC:%.*]] = icmp eq i64 [[IV_NEXT]], 100
; CHECK-NEXT: br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: [[EXIT]]:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%arrayidx = getelementptr inbounds i64, ptr %factor, i64 %iv
%l.factor = load i64, ptr %arrayidx, align 8
%1 = shl nsw i64 %iv, 1
%data.0 = getelementptr inbounds i64, ptr %data, i64 %1
%l.0 = load i64, ptr %data.0, align 8
%mul.0 = mul i64 %l.factor, %l.0
store i64 %mul.0, ptr %data.0, align 8
%3 = or disjoint i64 %1, 1
%data.1 = getelementptr inbounds i64, ptr %data, i64 %3
%l.1 = load i64, ptr %data.1, align 8
%mul.1 = mul i64 %l.factor, %l.1
store i64 %mul.1, ptr %data.1, align 8
%iv.next = add nuw nsw i64 %iv, 1
%ec = icmp eq i64 %iv.next, 100
br i1 %ec, label %exit, label %loop
exit:
ret void
}
|
