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
|
; RUN: opt -S -loop-predication < %s 2>&1 | FileCheck %s
; RUN: opt -S -passes='require<scalar-evolution>,loop(loop-predication)' < %s 2>&1 | FileCheck %s
declare void @llvm.experimental.guard(i1, ...)
define i32 @signed_loop_0_to_n_nested_0_to_l_inner_index_check(i32* %array, i32 %length, i32 %n, i32 %l) {
; CHECK-LABEL: @signed_loop_0_to_n_nested_0_to_l_inner_index_check
entry:
%tmp5 = icmp sle i32 %n, 0
br i1 %tmp5, label %exit, label %outer.loop.preheader
outer.loop.preheader:
br label %outer.loop
outer.loop:
%outer.loop.acc = phi i32 [ %outer.loop.acc.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%i = phi i32 [ %i.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%tmp6 = icmp sle i32 %l, 0
br i1 %tmp6, label %outer.loop.inc, label %inner.loop.preheader
inner.loop.preheader:
; CHECK: inner.loop.preheader:
; CHECK: [[limit_check:[^ ]+]] = icmp sle i32 %l, %length
; CHECK-NEXT: [[first_iteration_check:[^ ]+]] = icmp ult i32 0, %length
; CHECK-NEXT: [[wide_cond:[^ ]+]] = and i1 [[first_iteration_check]], [[limit_check]]
; CHECK-NEXT: br label %inner.loop
br label %inner.loop
inner.loop:
; CHECK: inner.loop:
; CHECK: call void (i1, ...) @llvm.experimental.guard(i1 [[wide_cond]], i32 9) [ "deopt"() ]
%inner.loop.acc = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %inner.loop.preheader ]
%j = phi i32 [ %j.next, %inner.loop ], [ 0, %inner.loop.preheader ]
%within.bounds = icmp ult i32 %j, %length
call void (i1, ...) @llvm.experimental.guard(i1 %within.bounds, i32 9) [ "deopt"() ]
%j.i64 = zext i32 %j to i64
%array.j.ptr = getelementptr inbounds i32, i32* %array, i64 %j.i64
%array.j = load i32, i32* %array.j.ptr, align 4
%inner.loop.acc.next = add i32 %inner.loop.acc, %array.j
%j.next = add nsw i32 %j, 1
%inner.continue = icmp slt i32 %j.next, %l
br i1 %inner.continue, label %inner.loop, label %outer.loop.inc
outer.loop.inc:
%outer.loop.acc.next = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %outer.loop ]
%i.next = add nsw i32 %i, 1
%outer.continue = icmp slt i32 %i.next, %n
br i1 %outer.continue, label %outer.loop, label %exit
exit:
%result = phi i32 [ 0, %entry ], [ %outer.loop.acc.next, %outer.loop.inc ]
ret i32 %result
}
define i32 @signed_loop_0_to_n_nested_0_to_l_outer_index_check(i32* %array, i32 %length, i32 %n, i32 %l) {
; CHECK-LABEL: @signed_loop_0_to_n_nested_0_to_l_outer_index_check
entry:
%tmp5 = icmp sle i32 %n, 0
br i1 %tmp5, label %exit, label %outer.loop.preheader
outer.loop.preheader:
; CHECK: outer.loop.preheader:
; CHECK: [[limit_check:[^ ]+]] = icmp sle i32 %n, %length
; CHECK-NEXT: [[first_iteration_check:[^ ]+]] = icmp ult i32 0, %length
; CHECK-NEXT: [[wide_cond:[^ ]+]] = and i1 [[first_iteration_check]], [[limit_check]]
; CHECK-NEXT: br label %outer.loop
br label %outer.loop
outer.loop:
%outer.loop.acc = phi i32 [ %outer.loop.acc.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%i = phi i32 [ %i.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%tmp6 = icmp sle i32 %l, 0
br i1 %tmp6, label %outer.loop.inc, label %inner.loop.preheader
inner.loop.preheader:
br label %inner.loop
inner.loop:
; CHECK: inner.loop:
; CHECK: call void (i1, ...) @llvm.experimental.guard(i1 [[wide_cond]], i32 9) [ "deopt"() ]
%inner.loop.acc = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %inner.loop.preheader ]
%j = phi i32 [ %j.next, %inner.loop ], [ 0, %inner.loop.preheader ]
%within.bounds = icmp ult i32 %i, %length
call void (i1, ...) @llvm.experimental.guard(i1 %within.bounds, i32 9) [ "deopt"() ]
%i.i64 = zext i32 %i to i64
%array.i.ptr = getelementptr inbounds i32, i32* %array, i64 %i.i64
%array.i = load i32, i32* %array.i.ptr, align 4
%inner.loop.acc.next = add i32 %inner.loop.acc, %array.i
%j.next = add nsw i32 %j, 1
%inner.continue = icmp slt i32 %j.next, %l
br i1 %inner.continue, label %inner.loop, label %outer.loop.inc
outer.loop.inc:
%outer.loop.acc.next = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %outer.loop ]
%i.next = add nsw i32 %i, 1
%outer.continue = icmp slt i32 %i.next, %n
br i1 %outer.continue, label %outer.loop, label %exit
exit:
%result = phi i32 [ 0, %entry ], [ %outer.loop.acc.next, %outer.loop.inc ]
ret i32 %result
}
define i32 @signed_loop_0_to_n_nested_i_to_l_inner_index_check(i32* %array, i32 %length, i32 %n, i32 %l) {
; CHECK-LABEL: @signed_loop_0_to_n_nested_i_to_l_inner_index_check
entry:
%tmp5 = icmp sle i32 %n, 0
br i1 %tmp5, label %exit, label %outer.loop.preheader
outer.loop.preheader:
; CHECK: outer.loop.preheader:
; CHECK-NEXT: [[limit_check_outer:[^ ]+]] = icmp sle i32 %n, %length
; CHECK-NEXT: [[first_iteration_check_outer:[^ ]+]] = icmp ult i32 0, %length
; CHECK-NEXT: [[wide_cond_outer:[^ ]+]] = and i1 [[first_iteration_check_outer]], [[limit_check_outer]]
; CHECK-NEXT: br label %outer.loop
br label %outer.loop
outer.loop:
; CHECK: outer.loop:
%outer.loop.acc = phi i32 [ %outer.loop.acc.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%i = phi i32 [ %i.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%tmp6 = icmp sle i32 %l, 0
br i1 %tmp6, label %outer.loop.inc, label %inner.loop.preheader
inner.loop.preheader:
; CHECK: inner.loop.preheader:
; CHECK: [[limit_check_inner:[^ ]+]] = icmp sle i32 %l, %length
; CHECK: br label %inner.loop
br label %inner.loop
inner.loop:
; CHECK: inner.loop:
; CHECK: [[wide_cond:[^ ]+]] = and i1 [[limit_check_inner]], [[wide_cond_outer]]
; CHECK: call void (i1, ...) @llvm.experimental.guard(i1 [[wide_cond]], i32 9) [ "deopt"() ]
%inner.loop.acc = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %inner.loop.preheader ]
%j = phi i32 [ %j.next, %inner.loop ], [ %i, %inner.loop.preheader ]
%within.bounds = icmp ult i32 %j, %length
call void (i1, ...) @llvm.experimental.guard(i1 %within.bounds, i32 9) [ "deopt"() ]
%j.i64 = zext i32 %j to i64
%array.j.ptr = getelementptr inbounds i32, i32* %array, i64 %j.i64
%array.j = load i32, i32* %array.j.ptr, align 4
%inner.loop.acc.next = add i32 %inner.loop.acc, %array.j
%j.next = add nsw i32 %j, 1
%inner.continue = icmp slt i32 %j.next, %l
br i1 %inner.continue, label %inner.loop, label %outer.loop.inc
outer.loop.inc:
%outer.loop.acc.next = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %outer.loop ]
%i.next = add nsw i32 %i, 1
%outer.continue = icmp slt i32 %i.next, %n
br i1 %outer.continue, label %outer.loop, label %exit
exit:
%result = phi i32 [ 0, %entry ], [ %outer.loop.acc.next, %outer.loop.inc ]
ret i32 %result
}
define i32 @cant_expand_guard_check_start(i32* %array, i32 %length, i32 %n, i32 %l, i32 %maybezero) {
; CHECK-LABEL: @cant_expand_guard_check_start
entry:
%tmp5 = icmp sle i32 %n, 0
br i1 %tmp5, label %exit, label %outer.loop.preheader
outer.loop.preheader:
br label %outer.loop
outer.loop:
%outer.loop.acc = phi i32 [ %outer.loop.acc.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%i = phi i32 [ %i.next, %outer.loop.inc ], [ 0, %outer.loop.preheader ]
%tmp6 = icmp sle i32 %l, 0
%div = udiv i32 %i, %maybezero
br i1 %tmp6, label %outer.loop.inc, label %inner.loop.preheader
inner.loop.preheader:
; CHECK: inner.loop.preheader:
; CHECK: br label %inner.loop
br label %inner.loop
inner.loop:
; CHECK: inner.loop:
; CHECK: %within.bounds = icmp ult i32 %j, %length
; CHECK: call void (i1, ...) @llvm.experimental.guard(i1 %within.bounds, i32 9) [ "deopt"() ]
%inner.loop.acc = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %inner.loop.preheader ]
%j = phi i32 [ %j.next, %inner.loop ], [ %div, %inner.loop.preheader ]
%within.bounds = icmp ult i32 %j, %length
call void (i1, ...) @llvm.experimental.guard(i1 %within.bounds, i32 9) [ "deopt"() ]
%j.i64 = zext i32 %j to i64
%array.j.ptr = getelementptr inbounds i32, i32* %array, i64 %j.i64
%array.j = load i32, i32* %array.j.ptr, align 4
%inner.loop.acc.next = add i32 %inner.loop.acc, %array.j
%j.next = add nsw i32 %j, 1
%inner.continue = icmp slt i32 %j.next, %l
br i1 %inner.continue, label %inner.loop, label %outer.loop.inc
outer.loop.inc:
%outer.loop.acc.next = phi i32 [ %inner.loop.acc.next, %inner.loop ], [ %outer.loop.acc, %outer.loop ]
%i.next = add nsw i32 %i, 1
%outer.continue = icmp slt i32 %i.next, %n
br i1 %outer.continue, label %outer.loop, label %exit
exit:
%result = phi i32 [ 0, %entry ], [ %outer.loop.acc.next, %outer.loop.inc ]
ret i32 %result
}
|
