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
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
|
; REQUIRES: x86-registered-target
; RUN: opt -aa-pipeline=basic-aa -passes=loop-distribute -enable-loop-distribute -verify-loop-info -verify-dom-info -S \
; RUN: < %s | FileCheck %s
; RUN: opt -aa-pipeline=basic-aa -passes='loop-distribute,loop-vectorize' -enable-loop-distribute -force-vector-width=4 \
; RUN: -verify-loop-info -verify-dom-info -S < %s | \
; RUN: FileCheck --check-prefix=VECTORIZE %s
; RUN: opt -aa-pipeline=basic-aa -passes='loop-distribute,print<access-info>' -enable-loop-distribute \
; RUN: -verify-loop-info -verify-dom-info -disable-output < %s 2>&1 | FileCheck %s --check-prefix=ANALYSIS
; The memcheck version of basic.ll. We should distribute and vectorize the
; second part of this loop with 5 memchecks (A+1 x {C, D, E} + C x {A, B})
;
; for (i = 0; i < n; i++) {
; A[i + 1] = A[i] * B[i];
; -------------------------------
; C[i] = D[i] * E[i];
; }
target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.10.0"
@B = common global ptr null, align 8
@A = common global ptr null, align 8
@C = common global ptr null, align 8
@D = common global ptr null, align 8
@E = common global ptr null, align 8
; CHECK-LABEL: @f(
define void @f() {
entry:
%a = load ptr, ptr @A, align 8
%b = load ptr, ptr @B, align 8
%c = load ptr, ptr @C, align 8
%d = load ptr, ptr @D, align 8
%e = load ptr, ptr @E, align 8
br label %for.body
; We have two compares for each array overlap check.
; Since the checks to A and A + 4 get merged, this will give us a
; total of 8 compares.
;
; CHECK: for.body.lver.check:
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK: = icmp
; CHECK-NOT: = icmp
; CHECK: br i1 %conflict.rdx15, label %for.body.ph.lver.orig, label %for.body.ph.ldist1
; The non-distributed loop that the memchecks fall back on.
; CHECK: for.body.ph.lver.orig:
; CHECK: br label %for.body.lver.orig
; CHECK: for.body.lver.orig:
; CHECK: br i1 %exitcond.lver.orig, label %for.end.loopexit, label %for.body.lver.orig
; Verify the two distributed loops.
; CHECK: for.body.ph.ldist1:
; CHECK: br label %for.body.ldist1
; CHECK: for.body.ldist1:
; CHECK: %mulA.ldist1 = mul i32 %loadB.ldist1, %loadA.ldist1
; CHECK: br i1 %exitcond.ldist1, label %for.body.ph, label %for.body.ldist1
; CHECK: for.body.ph:
; CHECK: br label %for.body
; CHECK: for.body:
; CHECK: %mulC = mul i32 %loadD, %loadE
; CHECK: for.end:
; VECTORIZE: mul <4 x i32>
; VECTORIZE: mul <4 x i32>
; VECTORIZE-NOT: mul <4 x i32>
for.body: ; preds = %for.body, %entry
%ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
%arrayidxA = getelementptr inbounds i32, ptr %a, i64 %ind
%loadA = load i32, ptr %arrayidxA, align 4
%arrayidxB = getelementptr inbounds i32, ptr %b, i64 %ind
%loadB = load i32, ptr %arrayidxB, align 4
%mulA = mul i32 %loadB, %loadA
%add = add nuw nsw i64 %ind, 1
%arrayidxA_plus_4 = getelementptr inbounds i32, ptr %a, i64 %add
store i32 %mulA, ptr %arrayidxA_plus_4, align 4
%arrayidxD = getelementptr inbounds i32, ptr %d, i64 %ind
%loadD = load i32, ptr %arrayidxD, align 4
%arrayidxE = getelementptr inbounds i32, ptr %e, i64 %ind
%loadE = load i32, ptr %arrayidxE, align 4
%mulC = mul i32 %loadD, %loadE
%arrayidxC = getelementptr inbounds i32, ptr %c, i64 %ind
store i32 %mulC, ptr %arrayidxC, align 4
%exitcond = icmp eq i64 %add, 20
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; Make sure there's no "Multiple reports generated" assert with a
; volatile load, and no distribution
; TODO: Distribution of volatile may be possible under some
; circumstance, but the current implementation does not touch them.
; CHECK-LABEL: @f_volatile_load(
; CHECK: br label %for.body{{$}}
; CHECK-NOT: load
; CHECK: {{^}}for.body:
; CHECK: load i32
; CHECK: load i32
; CHECK: load volatile i32
; CHECK: load i32
; CHECK: br i1 %exitcond, label %for.end, label %for.body{{$}}
; CHECK-NOT: load
; VECTORIZE-NOT: load <4 x i32>
; VECTORIZE-NOT: mul <4 x i32>
define void @f_volatile_load() {
entry:
%a = load ptr, ptr @A, align 8
%b = load ptr, ptr @B, align 8
%c = load ptr, ptr @C, align 8
%d = load ptr, ptr @D, align 8
%e = load ptr, ptr @E, align 8
br label %for.body
for.body:
%ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
%arrayidxA = getelementptr inbounds i32, ptr %a, i64 %ind
%loadA = load i32, ptr %arrayidxA, align 4
%arrayidxB = getelementptr inbounds i32, ptr %b, i64 %ind
%loadB = load i32, ptr %arrayidxB, align 4
%mulA = mul i32 %loadB, %loadA
%add = add nuw nsw i64 %ind, 1
%arrayidxA_plus_4 = getelementptr inbounds i32, ptr %a, i64 %add
store i32 %mulA, ptr %arrayidxA_plus_4, align 4
%arrayidxD = getelementptr inbounds i32, ptr %d, i64 %ind
%loadD = load volatile i32, ptr %arrayidxD, align 4
%arrayidxE = getelementptr inbounds i32, ptr %e, i64 %ind
%loadE = load i32, ptr %arrayidxE, align 4
%mulC = mul i32 %loadD, %loadE
%arrayidxC = getelementptr inbounds i32, ptr %c, i64 %ind
store i32 %mulC, ptr %arrayidxC, align 4
%exitcond = icmp eq i64 %add, 20
br i1 %exitcond, label %for.end, label %for.body
for.end:
ret void
}
declare i32 @llvm.convergent(i32) #0
; This is the same as f, and would require the same bounds
; check. However, it is not OK to introduce new control dependencies
; on the convergent call.
; CHECK-LABEL: @f_with_convergent(
; CHECK: call i32 @llvm.convergent
; CHECK-NOT: call i32 @llvm.convergent
; ANALYSIS: for.body:
; ANALYSIS: Report: cannot add control dependency to convergent operation
define void @f_with_convergent() #1 {
entry:
%a = load ptr, ptr @A, align 8
%b = load ptr, ptr @B, align 8
%c = load ptr, ptr @C, align 8
%d = load ptr, ptr @D, align 8
%e = load ptr, ptr @E, align 8
br label %for.body
for.body: ; preds = %for.body, %entry
%ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
%arrayidxA = getelementptr inbounds i32, ptr %a, i64 %ind
%loadA = load i32, ptr %arrayidxA, align 4
%arrayidxB = getelementptr inbounds i32, ptr %b, i64 %ind
%loadB = load i32, ptr %arrayidxB, align 4
%mulA = mul i32 %loadB, %loadA
%add = add nuw nsw i64 %ind, 1
%arrayidxA_plus_4 = getelementptr inbounds i32, ptr %a, i64 %add
store i32 %mulA, ptr %arrayidxA_plus_4, align 4
%arrayidxD = getelementptr inbounds i32, ptr %d, i64 %ind
%loadD = load i32, ptr %arrayidxD, align 4
%arrayidxE = getelementptr inbounds i32, ptr %e, i64 %ind
%loadE = load i32, ptr %arrayidxE, align 4
%convergentD = call i32 @llvm.convergent(i32 %loadD)
%mulC = mul i32 %convergentD, %loadE
%arrayidxC = getelementptr inbounds i32, ptr %c, i64 %ind
store i32 %mulC, ptr %arrayidxC, align 4
%exitcond = icmp eq i64 %add, 20
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; Make sure an explicit request for distribution is ignored if it
; requires possibly illegal checks.
; CHECK-LABEL: @f_with_convergent_forced_distribute(
; CHECK: call i32 @llvm.convergent
; CHECK-NOT: call i32 @llvm.convergent
define void @f_with_convergent_forced_distribute() #1 {
entry:
%a = load ptr, ptr @A, align 8
%b = load ptr, ptr @B, align 8
%c = load ptr, ptr @C, align 8
%d = load ptr, ptr @D, align 8
%e = load ptr, ptr @E, align 8
br label %for.body
for.body: ; preds = %for.body, %entry
%ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
%arrayidxA = getelementptr inbounds i32, ptr %a, i64 %ind
%loadA = load i32, ptr %arrayidxA, align 4
%arrayidxB = getelementptr inbounds i32, ptr %b, i64 %ind
%loadB = load i32, ptr %arrayidxB, align 4
%mulA = mul i32 %loadB, %loadA
%add = add nuw nsw i64 %ind, 1
%arrayidxA_plus_4 = getelementptr inbounds i32, ptr %a, i64 %add
store i32 %mulA, ptr %arrayidxA_plus_4, align 4
%arrayidxD = getelementptr inbounds i32, ptr %d, i64 %ind
%loadD = load i32, ptr %arrayidxD, align 4
%arrayidxE = getelementptr inbounds i32, ptr %e, i64 %ind
%loadE = load i32, ptr %arrayidxE, align 4
%convergentD = call i32 @llvm.convergent(i32 %loadD)
%mulC = mul i32 %convergentD, %loadE
%arrayidxC = getelementptr inbounds i32, ptr %c, i64 %ind
store i32 %mulC, ptr %arrayidxC, align 4
%exitcond = icmp eq i64 %add, 20
br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !0
for.end: ; preds = %for.body
ret void
}
attributes #0 = { nounwind readnone convergent }
attributes #1 = { nounwind convergent }
!0 = distinct !{!0, !1}
!1 = !{!"llvm.loop.distribute.enable", i1 true}
|
