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
|
; REQUIRES: asserts
; RUN: opt -mattr=+neon,+dotprod -passes=loop-vectorize -debug-only=loop-vectorize -force-vector-interleave=1 -enable-epilogue-vectorization -epilogue-vectorization-force-VF=2 -disable-output %s 2>&1 | FileCheck %s
target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
target triple = "aarch64-none-unknown-elf"
; Tests for printing VPlans that are enabled under AArch64
define i32 @print_partial_reduction(ptr %a, ptr %b) {
; CHECK: VPlan 'Initial VPlan for VF={8,16},UF>=1' {
; CHECK-NEXT: Live-in vp<[[VFxUF:%.]]> = VF * UF
; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
; CHECK-NEXT: Live-in ir<1024> = original trip-count
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<entry>:
; CHECK-NEXT: Successor(s): vector.ph
; CHECK-EMPTY:
; CHECK-NEXT: vector.ph:
; CHECK-NEXT: Successor(s): vector loop
; CHECK-EMPTY:
; CHECK-NEXT: <x1> vector loop: {
; CHECK-NEXT: vector.body:
; CHECK-NEXT: EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION ir<0>, vp<[[CAN_IV_NEXT:%.+]]>
; CHECK-NEXT: WIDEN-REDUCTION-PHI ir<[[ACC:%.+]]> = phi ir<0>, ir<[[REDUCE:%.+]]> (VF scaled by 1/4)
; CHECK-NEXT: vp<[[STEPS:%.+]]> = SCALAR-STEPS vp<[[CAN_IV]]>, ir<1>
; CHECK-NEXT: CLONE ir<%gep.a> = getelementptr ir<%a>, vp<[[STEPS]]>
; CHECK-NEXT: vp<[[PTR_A:%.+]]> = vector-pointer ir<%gep.a>
; CHECK-NEXT: WIDEN ir<%load.a> = load vp<[[PTR_A]]>
; CHECK-NEXT: WIDEN-CAST ir<%ext.a> = zext ir<%load.a> to i32
; CHECK-NEXT: CLONE ir<%gep.b> = getelementptr ir<%b>, vp<[[STEPS]]>
; CHECK-NEXT: vp<[[PTR_B:%.+]]> = vector-pointer ir<%gep.b>
; CHECK-NEXT: WIDEN ir<%load.b> = load vp<[[PTR_B]]>
; CHECK-NEXT: WIDEN-CAST ir<%ext.b> = zext ir<%load.b> to i32
; CHECK-NEXT: WIDEN ir<%mul> = mul ir<%ext.b>, ir<%ext.a>
; CHECK-NEXT: PARTIAL-REDUCE ir<[[REDUCE]]> = add ir<%mul>, ir<[[ACC]]>
; CHECK-NEXT: EMIT vp<[[CAN_IV_NEXT]]> = add nuw vp<[[CAN_IV]]>, vp<[[VFxUF]]>
; CHECK-NEXT: EMIT branch-on-count vp<[[CAN_IV_NEXT]]>, vp<[[VEC_TC]]>
; CHECK-NEXT: No successors
; CHECK-NEXT: }
; CHECK-NEXT: Successor(s): middle.block
; CHECK-EMPTY:
; CHECK-NEXT: middle.block:
; CHECK-NEXT: EMIT vp<[[RED_RESULT:%.+]]> = compute-reduction-result ir<[[ACC]]>, ir<[[REDUCE]]>
; CHECK-NEXT: EMIT vp<[[EXTRACT:%.+]]> = extract-from-end vp<[[RED_RESULT]]>, ir<1>
; CHECK-NEXT: EMIT vp<[[CMP:%.+]]> = icmp eq ir<1024>, vp<%1>
; CHECK-NEXT: EMIT branch-on-cond vp<[[CMP]]>
; CHECK-NEXT: Successor(s): ir-bb<exit>, scalar.ph
; CHECK-EMPTY:
; CHECK-NEXT: scalar.ph:
; CHECK-NEXT: EMIT vp<%bc.resume.val> = resume-phi vp<[[VEC_TC]]>, ir<0>
; CHECK-NEXT: EMIT vp<%bc.merge.rdx> = resume-phi vp<[[RED_RESULT]]>, ir<0>
; CHECK-NEXT: Successor(s): ir-bb<for.body>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<for.body>:
; CHECK-NEXT: IR %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
; CHECK-NEXT: IR %accum = phi i32 [ 0, %entry ], [ %add, %for.body ] (extra operand: vp<%bc.merge.rdx> from scalar.ph)
; CHECK-NEXT: IR %gep.a = getelementptr i8, ptr %a, i64 %iv
; CHECK-NEXT: IR %load.a = load i8, ptr %gep.a, align 1
; CHECK-NEXT: IR %ext.a = zext i8 %load.a to i32
; CHECK-NEXT: IR %gep.b = getelementptr i8, ptr %b, i64 %iv
; CHECK-NEXT: IR %load.b = load i8, ptr %gep.b, align 1
; CHECK-NEXT: IR %ext.b = zext i8 %load.b to i32
; CHECK-NEXT: IR %mul = mul i32 %ext.b, %ext.a
; CHECK-NEXT: IR %add = add i32 %mul, %accum
; CHECK-NEXT: IR %iv.next = add i64 %iv, 1
; CHECK-NEXT: IR %exitcond.not = icmp eq i64 %iv.next, 1024
; CHECK-NEXT: No successors
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<exit>:
; CHECK-NEXT: IR %add.lcssa = phi i32 [ %add, %for.body ] (extra operand: vp<[[EXTRACT]]> from middle.block)
; CHECK-NEXT: No successors
; CHECK-NEXT: }
; CHECK: VPlan 'Final VPlan for VF={8,16},UF={1}' {
; CHECK-NEXT: Live-in ir<[[EP_VFxUF:.+]]> = VF * UF
; CHECK-NEXT: Live-in ir<[[EP_VEC_TC:.+]]> = vector-trip-count
; CHECK-NEXT: Live-in ir<1024> = original trip-count
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<entry>:
; CHECK-NEXT: Successor(s): ir-bb<scalar.ph>, ir-bb<vector.main.loop.iter.check>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<vector.main.loop.iter.check>:
; CHECK-NEXT: Successor(s): ir-bb<scalar.ph>, ir-bb<vector.ph>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<vector.ph>:
; CHECK-NEXT: Successor(s): vector loop
; CHECK-EMPTY:
; CHECK-NEXT: <x1> vector loop: {
; CHECK-NEXT: vector.body:
; CHECK-NEXT: SCALAR-PHI vp<[[EP_IV:%.+]]> = phi ir<0>, vp<%index.next>
; CHECK-NEXT: WIDEN-REDUCTION-PHI ir<%accum> = phi ir<0>, ir<%add> (VF scaled by 1/4)
; CHECK-NEXT: vp<[[STEPS:%.+]]> = SCALAR-STEPS vp<[[EP_IV]]>, ir<1>
; CHECK-NEXT: CLONE ir<%gep.a> = getelementptr ir<%a>, vp<[[STEPS]]>
; CHECK-NEXT: vp<[[PTR_A:%.+]]> = vector-pointer ir<%gep.a>
; CHECK-NEXT: WIDEN ir<%load.a> = load vp<[[PTR_A]]>
; CHECK-NEXT: WIDEN-CAST ir<%ext.a> = zext ir<%load.a> to i32
; CHECK-NEXT: CLONE ir<%gep.b> = getelementptr ir<%b>, vp<[[STEPS]]>
; CHECK-NEXT: vp<[[PTR_B:%.+]]> = vector-pointer ir<%gep.b>
; CHECK-NEXT: WIDEN ir<%load.b> = load vp<[[PTR_B]]>
; CHECK-NEXT: WIDEN-CAST ir<%ext.b> = zext ir<%load.b> to i32
; CHECK-NEXT: WIDEN ir<%mul> = mul ir<%ext.b>, ir<%ext.a>
; CHECK-NEXT: PARTIAL-REDUCE ir<%add> = add ir<%mul>, ir<%accum>
; CHECK-NEXT: EMIT vp<[[EP_IV_NEXT:%.+]]> = add nuw vp<[[EP_IV]]>, ir<16>
; CHECK-NEXT: EMIT branch-on-count vp<[[EP_IV_NEXT]]>, ir<1024>
; CHECK-NEXT: No successors
; CHECK-NEXT: }
; CHECK-NEXT: Successor(s): ir-bb<middle.block>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<middle.block>:
; CHECK-NEXT: EMIT vp<[[RED_RESULT:%.+]]> = compute-reduction-result ir<%accum>, ir<%add>
; CHECK-NEXT: EMIT vp<[[EXTRACT:%.+]]> = extract-from-end vp<[[RED_RESULT]]>, ir<1>
; CHECK-NEXT: EMIT vp<[[CMP:%.+]]> = icmp eq ir<1024>, ir<1024>
; CHECK-NEXT: EMIT branch-on-cond vp<[[CMP]]>
; CHECK-NEXT: Successor(s): ir-bb<exit>, ir-bb<scalar.ph>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<exit>:
; CHECK-NEXT: IR %add.lcssa = phi i32 [ %add, %for.body ] (extra operand: vp<[[EXTRACT]]> from ir-bb<middle.block>)
; CHECK-NEXT: No successors
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<scalar.ph>:
; CHECK-NEXT: EMIT vp<[[EP_RESUME:%.+]]> = resume-phi ir<1024>, ir<0>
; CHECK-NEXT: EMIT vp<[[EP_MERGE:%.+]]> = resume-phi vp<[[RED_RESULT]]>, ir<0>
; CHECK-NEXT: Successor(s): ir-bb<for.body>
; CHECK-EMPTY:
; CHECK-NEXT: ir-bb<for.body>:
; CHECK-NEXT: IR %accum = phi i32 [ 0, %scalar.ph ], [ %add, %for.body ] (extra operand: vp<[[EP_MERGE]]> from ir-bb<scalar.ph>)
; CHECK-NEXT: IR %gep.a = getelementptr i8, ptr %a, i64 %iv
; CHECK-NEXT: IR %load.a = load i8, ptr %gep.a, align 1
; CHECK-NEXT: IR %ext.a = zext i8 %load.a to i32
; CHECK-NEXT: IR %gep.b = getelementptr i8, ptr %b, i64 %iv
; CHECK-NEXT: IR %load.b = load i8, ptr %gep.b, align 1
; CHECK-NEXT: IR %ext.b = zext i8 %load.b to i32
; CHECK-NEXT: IR %mul = mul i32 %ext.b, %ext.a
; CHECK-NEXT: IR %add = add i32 %mul, %accum
; CHECK-NEXT: IR %iv.next = add i64 %iv, 1
; CHECK-NEXT: IR %exitcond.not = icmp eq i64 %iv.next, 1024
; CHECK-NEXT: No successors
; CHECK-NEXT: }
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
%accum = phi i32 [ 0, %entry ], [ %add, %for.body ]
%gep.a = getelementptr i8, ptr %a, i64 %iv
%load.a = load i8, ptr %gep.a, align 1
%ext.a = zext i8 %load.a to i32
%gep.b = getelementptr i8, ptr %b, i64 %iv
%load.b = load i8, ptr %gep.b, align 1
%ext.b = zext i8 %load.b to i32
%mul = mul i32 %ext.b, %ext.a
%add = add i32 %mul, %accum
%iv.next = add i64 %iv, 1
%exitcond.not = icmp eq i64 %iv.next, 1024
br i1 %exitcond.not, label %exit, label %for.body
exit:
ret i32 %add
}
|
