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
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
|
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 4
; RUN: opt -passes='print<access-info>' -disable-output %s 2>&1 | FileCheck %s
target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128"
declare void @llvm.assume(i1)
declare void @use(ptr noundef)
; %offset is known positive via assume, so we should be able to detect the
; forward dependence.
define void @offset_i8_known_positive_via_assume_forward_dep_1(ptr %A, i64 %offset, i64 %N) {
; CHECK-LABEL: 'offset_i8_known_positive_via_assume_forward_dep_1'
; CHECK-NEXT: loop:
; CHECK-NEXT: Memory dependences are safe
; CHECK-NEXT: Dependences:
; CHECK-NEXT: Forward:
; CHECK-NEXT: %l = load i8, ptr %gep.off, align 4 ->
; CHECK-NEXT: store i8 %add, ptr %gep, align 4
; CHECK-EMPTY:
; CHECK-NEXT: Run-time memory checks:
; CHECK-NEXT: Grouped accesses:
; CHECK-EMPTY:
; CHECK-NEXT: Non vectorizable stores to invariant address were not found in loop.
; CHECK-NEXT: SCEV assumptions:
; CHECK-EMPTY:
; CHECK-NEXT: Expressions re-written:
;
entry:
%c = icmp sgt i64 %offset, 0
call void @llvm.assume(i1 %c)
%off = getelementptr inbounds i8, ptr %A, i64 %offset
call void @use(ptr noundef %off)
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.off = getelementptr inbounds i8, ptr %off, i64 %iv
%l = load i8 , ptr %gep.off, align 4
%add = add nsw i8 %l, 5
%gep = getelementptr inbounds i8, ptr %A, i64 %iv
store i8 %add, ptr %gep, align 4
%iv.next = add nuw nsw i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, %N
br i1 %exitcond.not, label %exit, label %loop
exit:
ret void
}
define void @offset_i32_known_positive_via_assume_forward_dep_1(ptr %A, i64 %offset, i64 %N) {
; CHECK-LABEL: 'offset_i32_known_positive_via_assume_forward_dep_1'
; CHECK-NEXT: loop:
; CHECK-NEXT: Memory dependences are safe with run-time checks
; CHECK-NEXT: Dependences:
; CHECK-NEXT: Run-time memory checks:
; CHECK-NEXT: Check 0:
; CHECK-NEXT: Comparing group GRP0:
; CHECK-NEXT: %gep = getelementptr inbounds i32, ptr %A, i64 %iv
; CHECK-NEXT: Against group GRP1:
; CHECK-NEXT: %gep.off = getelementptr inbounds i32, ptr %off, i64 %iv
; CHECK-NEXT: Grouped accesses:
; CHECK-NEXT: Group GRP0:
; CHECK-NEXT: (Low: %A High: (-3 + (4 * %N) + %A))
; CHECK-NEXT: Member: {%A,+,4}<nuw><%loop>
; CHECK-NEXT: Group GRP1:
; CHECK-NEXT: (Low: ((4 * %offset)<nsw> + %A) High: (-3 + (4 * %offset)<nsw> + (4 * %N) + %A))
; CHECK-NEXT: Member: {((4 * %offset)<nsw> + %A),+,4}<nw><%loop>
; CHECK-EMPTY:
; CHECK-NEXT: Non vectorizable stores to invariant address were not found in loop.
; CHECK-NEXT: SCEV assumptions:
; CHECK-EMPTY:
; CHECK-NEXT: Expressions re-written:
;
entry:
%c = icmp sgt i64 %offset, 0
call void @llvm.assume(i1 %c)
%off = getelementptr inbounds i32, ptr %A, i64 %offset
call void @use(ptr noundef %off)
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.off = getelementptr inbounds i32, ptr %off, i64 %iv
%l = load i8 , ptr %gep.off, align 4
%add = add nsw i8 %l, 5
%gep = getelementptr inbounds i32, ptr %A, i64 %iv
store i8 %add, ptr %gep, align 4
%iv.next = add nuw nsw i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, %N
br i1 %exitcond.not, label %exit, label %loop
exit:
ret void
}
; %offset is known positive via assume, so we should be able to detect the
; forward dependence.
define void @offset_known_positive_via_assume_forward_dep_2(ptr %A, i64 %offset, i64 %N) {
; CHECK-LABEL: 'offset_known_positive_via_assume_forward_dep_2'
; CHECK-NEXT: loop:
; CHECK-NEXT: Memory dependences are safe
; CHECK-NEXT: Dependences:
; CHECK-NEXT: Forward:
; CHECK-NEXT: %l = load i32, ptr %gep.off, align 4 ->
; CHECK-NEXT: store i32 %add, ptr %gep, align 4
; CHECK-EMPTY:
; CHECK-NEXT: Run-time memory checks:
; CHECK-NEXT: Grouped accesses:
; CHECK-EMPTY:
; CHECK-NEXT: Non vectorizable stores to invariant address were not found in loop.
; CHECK-NEXT: SCEV assumptions:
; CHECK-EMPTY:
; CHECK-NEXT: Expressions re-written:
;
entry:
%c = icmp sgt i64 %offset, 0
call void @llvm.assume(i1 %c)
%c.2 = icmp slt i64 %offset, 20
call void @llvm.assume(i1 %c.2)
%off = getelementptr inbounds i32, ptr %A, i64 %offset
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.off = getelementptr inbounds i32, ptr %off, i64 %iv
%l = load i32, ptr %gep.off, align 4
%add = add nsw i32 %l, 5
%gep = getelementptr inbounds i32, ptr %A, i64 %iv
store i32 %add, ptr %gep, align 4
%iv.next = add nuw nsw i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, %N
br i1 %exitcond.not, label %exit, label %loop
exit:
ret void
}
; The range of %offset is known via assumes, but it may be positive or negative.
define void @offset_may_be_negative_via_assume_unknown_dep(ptr %A, i64 %offset, i64 %N) {
; CHECK-LABEL: 'offset_may_be_negative_via_assume_unknown_dep'
; CHECK-NEXT: loop:
; CHECK-NEXT: Memory dependences are safe with run-time checks
; CHECK-NEXT: Dependences:
; CHECK-NEXT: Run-time memory checks:
; CHECK-NEXT: Check 0:
; CHECK-NEXT: Comparing group GRP0:
; CHECK-NEXT: %gep.mul.2 = getelementptr inbounds i32, ptr %A, i64 %iv
; CHECK-NEXT: Against group GRP1:
; CHECK-NEXT: %gep = getelementptr inbounds i32, ptr %off, i64 %iv
; CHECK-NEXT: Grouped accesses:
; CHECK-NEXT: Group GRP0:
; CHECK-NEXT: (Low: %A High: ((4 * %N) + %A))
; CHECK-NEXT: Member: {%A,+,4}<nuw><%loop>
; CHECK-NEXT: Group GRP1:
; CHECK-NEXT: (Low: ((4 * %offset)<nsw> + %A) High: ((4 * %offset)<nsw> + (4 * %N) + %A))
; CHECK-NEXT: Member: {((4 * %offset)<nsw> + %A),+,4}<nw><%loop>
; CHECK-EMPTY:
; CHECK-NEXT: Non vectorizable stores to invariant address were not found in loop.
; CHECK-NEXT: SCEV assumptions:
; CHECK-EMPTY:
; CHECK-NEXT: Expressions re-written:
;
entry:
%c = icmp sgt i64 %offset, -4
call void @llvm.assume(i1 %c)
%c.2 = icmp slt i64 %offset, 20
call void @llvm.assume(i1 %c.2)
%off = getelementptr inbounds i32, ptr %A, i64 %offset
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep = getelementptr inbounds i32, ptr %off, i64 %iv
%l = load i32, ptr %gep, align 4
%add = add nsw i32 %l, 5
%gep.mul.2 = getelementptr inbounds i32, ptr %A, i64 %iv
store i32 %add, ptr %gep.mul.2, align 4
%iv.next = add nuw nsw i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, %N
br i1 %exitcond.not, label %exit, label %loop
exit:
ret void
}
define void @offset_no_assumes(ptr %A, i64 %offset, i64 %N) {
; CHECK-LABEL: 'offset_no_assumes'
; CHECK-NEXT: loop:
; CHECK-NEXT: Memory dependences are safe with run-time checks
; CHECK-NEXT: Dependences:
; CHECK-NEXT: Run-time memory checks:
; CHECK-NEXT: Check 0:
; CHECK-NEXT: Comparing group GRP0:
; CHECK-NEXT: %gep = getelementptr inbounds i32, ptr %A, i64 %iv
; CHECK-NEXT: Against group GRP1:
; CHECK-NEXT: %gep.off = getelementptr inbounds i32, ptr %off, i64 %iv
; CHECK-NEXT: Grouped accesses:
; CHECK-NEXT: Group GRP0:
; CHECK-NEXT: (Low: %A High: ((4 * %N) + %A))
; CHECK-NEXT: Member: {%A,+,4}<nuw><%loop>
; CHECK-NEXT: Group GRP1:
; CHECK-NEXT: (Low: ((4 * %offset)<nsw> + %A) High: ((4 * %offset)<nsw> + (4 * %N) + %A))
; CHECK-NEXT: Member: {((4 * %offset)<nsw> + %A),+,4}<nw><%loop>
; CHECK-EMPTY:
; CHECK-NEXT: Non vectorizable stores to invariant address were not found in loop.
; CHECK-NEXT: SCEV assumptions:
; CHECK-EMPTY:
; CHECK-NEXT: Expressions re-written:
;
entry:
%off = getelementptr inbounds i32, ptr %A, i64 %offset
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.off = getelementptr inbounds i32, ptr %off, i64 %iv
%l = load i32, ptr %gep.off, align 4
%add = add nsw i32 %l, 5
%gep = getelementptr inbounds i32, ptr %A, i64 %iv
store i32 %add, ptr %gep, align 4
%iv.next = add nuw nsw i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, %N
br i1 %exitcond.not, label %exit, label %loop
exit:
ret void
}
|
