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
|
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=thumbv8.1m.main-none-eabi -mattr=+lob,+mve -run-pass=arm-mve-vpt-opts %s -verify-machineinstrs -o - | FileCheck %s
--- |
define i32 @test(i16* nocapture readonly %x, i16* nocapture readonly %y, i32 %n) {
entry:
%cmp10 = icmp sgt i32 %n, 0
%0 = add i32 %n, 7
%1 = lshr i32 %0, 3
%2 = shl nuw i32 %1, 3
%3 = add i32 %2, -8
%4 = lshr i32 %3, 3
%5 = add nuw nsw i32 %4, 1
br i1 %cmp10, label %vector.ph, label %for.cond.cleanup
vector.ph: ; preds = %entry
%6 = call i32 @llvm.start.loop.iterations.i32(i32 %5)
br label %vector.body
vector.body: ; preds = %vector.body, %vector.ph
%lsr.iv3 = phi i16* [ %scevgep4, %vector.body ], [ %x, %vector.ph ]
%lsr.iv1 = phi i16* [ %scevgep, %vector.body ], [ %y, %vector.ph ]
%vec.phi = phi i32 [ 0, %vector.ph ], [ %16, %vector.body ]
%7 = phi i32 [ %6, %vector.ph ], [ %17, %vector.body ]
%8 = phi i32 [ %n, %vector.ph ], [ %10, %vector.body ]
%lsr.iv12 = bitcast i16* %lsr.iv1 to <8 x i16>*
%lsr.iv35 = bitcast i16* %lsr.iv3 to <8 x i16>*
%9 = call <8 x i1> @llvm.arm.mve.vctp16(i32 %8)
%10 = sub i32 %8, 8
%wide.masked.load = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %lsr.iv35, i32 2, <8 x i1> %9, <8 x i16> undef)
%11 = sext <8 x i16> %wide.masked.load to <8 x i32>
%wide.masked.load13 = call <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>* %lsr.iv12, i32 2, <8 x i1> %9, <8 x i16> undef)
%12 = sext <8 x i16> %wide.masked.load13 to <8 x i32>
%13 = mul nsw <8 x i32> %12, %11
%14 = select <8 x i1> %9, <8 x i32> %13, <8 x i32> zeroinitializer
%15 = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %14)
%16 = add i32 %15, %vec.phi
%scevgep = getelementptr i16, i16* %lsr.iv1, i32 8
%scevgep4 = getelementptr i16, i16* %lsr.iv3, i32 8
%17 = call i32 @llvm.loop.decrement.reg.i32(i32 %7, i32 1)
%18 = icmp ne i32 %17, 0
br i1 %18, label %vector.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %vector.body, %entry
%s.0.lcssa = phi i32 [ 0, %entry ], [ %16, %vector.body ]
ret i32 %s.0.lcssa
}
declare <8 x i1> @llvm.get.active.lane.mask.v8i1.i32(i32, i32)
declare <8 x i16> @llvm.masked.load.v8i16.p0v8i16(<8 x i16>*, i32 immarg, <8 x i1>, <8 x i16>)
declare i32 @llvm.vector.reduce.add.v8i32(<8 x i32>)
declare i32 @llvm.start.loop.iterations.i32(i32)
declare i32 @llvm.loop.decrement.reg.i32(i32, i32)
declare <8 x i1> @llvm.arm.mve.vctp16(i32)
...
---
name: test
alignment: 2
tracksRegLiveness: true
registers:
- { id: 0, class: rgpr, preferred-register: '' }
- { id: 1, class: gpr, preferred-register: '' }
- { id: 2, class: rgpr, preferred-register: '' }
- { id: 3, class: rgpr, preferred-register: '' }
- { id: 4, class: tgpreven, preferred-register: '' }
- { id: 5, class: gprlr, preferred-register: '' }
- { id: 6, class: rgpr, preferred-register: '' }
- { id: 7, class: gpr, preferred-register: '' }
- { id: 8, class: gpr, preferred-register: '' }
- { id: 9, class: gpr, preferred-register: '' }
- { id: 10, class: gpr, preferred-register: '' }
- { id: 11, class: gpr, preferred-register: '' }
- { id: 12, class: gpr, preferred-register: '' }
- { id: 13, class: gpr, preferred-register: '' }
- { id: 14, class: gpr, preferred-register: '' }
- { id: 15, class: gprnopc, preferred-register: '' }
- { id: 16, class: gpr, preferred-register: '' }
- { id: 17, class: rgpr, preferred-register: '' }
- { id: 18, class: rgpr, preferred-register: '' }
- { id: 19, class: rgpr, preferred-register: '' }
- { id: 20, class: rgpr, preferred-register: '' }
- { id: 21, class: gprnopc, preferred-register: '' }
- { id: 22, class: rgpr, preferred-register: '' }
- { id: 23, class: gpr, preferred-register: '' }
- { id: 24, class: gprlr, preferred-register: '' }
- { id: 25, class: rgpr, preferred-register: '' }
- { id: 26, class: vccr, preferred-register: '' }
- { id: 27, class: rgpr, preferred-register: '' }
- { id: 28, class: rgpr, preferred-register: '' }
- { id: 29, class: mqpr, preferred-register: '' }
- { id: 30, class: rgpr, preferred-register: '' }
- { id: 31, class: mqpr, preferred-register: '' }
- { id: 32, class: tgpreven, preferred-register: '' }
- { id: 33, class: gprlr, preferred-register: '' }
- { id: 34, class: gprlr, preferred-register: '' }
- { id: 35, class: gprnopc, preferred-register: '' }
liveins:
- { reg: '$r0', virtual-reg: '%13' }
- { reg: '$r1', virtual-reg: '%14' }
- { reg: '$r2', virtual-reg: '%15' }
body: |
; CHECK-LABEL: name: test
; CHECK: bb.0.entry:
; CHECK: successors: %bb.2(0x50000000), %bb.1(0x30000000)
; CHECK: liveins: $r0, $r1, $r2
; CHECK: [[COPY:%[0-9]+]]:gprnopc = COPY $r2
; CHECK: [[COPY1:%[0-9]+]]:gpr = COPY $r1
; CHECK: [[COPY2:%[0-9]+]]:gpr = COPY $r0
; CHECK: t2CMPri [[COPY]], 1, 14 /* CC::al */, $noreg, implicit-def $cpsr
; CHECK: t2Bcc %bb.2, 10 /* CC::ge */, $cpsr
; CHECK: bb.1:
; CHECK: successors: %bb.4(0x80000000)
; CHECK: [[t2MOVi:%[0-9]+]]:rgpr = t2MOVi 0, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[COPY3:%[0-9]+]]:gpr = COPY [[t2MOVi]]
; CHECK: t2B %bb.4, 14 /* CC::al */, $noreg
; CHECK: bb.2.vector.ph:
; CHECK: successors: %bb.3(0x80000000)
; CHECK: [[t2ADDri:%[0-9]+]]:rgpr = t2ADDri [[COPY]], 7, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[t2BICri:%[0-9]+]]:rgpr = t2BICri [[t2ADDri]], 7, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[t2SUBri:%[0-9]+]]:rgpr = t2SUBri [[t2BICri]], 8, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[t2MOVi1:%[0-9]+]]:rgpr = t2MOVi 1, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[t2ADDrs:%[0-9]+]]:gprnopc = nuw nsw t2ADDrs [[t2MOVi1]], [[t2SUBri]], 27, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[COPY4:%[0-9]+]]:rgpr = COPY [[t2ADDrs]]
; CHECK: [[t2MOVi2:%[0-9]+]]:rgpr = t2MOVi 0, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[COPY5:%[0-9]+]]:gpr = COPY [[t2MOVi2]]
; CHECK: [[COPY6:%[0-9]+]]:rgpr = COPY [[COPY]]
; CHECK: [[t2DoLoopStartTP:%[0-9]+]]:gprlr = t2DoLoopStartTP [[COPY4]], [[COPY6]]
; CHECK: bb.3.vector.body:
; CHECK: successors: %bb.3(0x7c000000), %bb.4(0x04000000)
; CHECK: [[PHI:%[0-9]+]]:rgpr = PHI [[COPY2]], %bb.2, %10, %bb.3
; CHECK: [[PHI1:%[0-9]+]]:rgpr = PHI [[COPY1]], %bb.2, %9, %bb.3
; CHECK: [[PHI2:%[0-9]+]]:tgpreven = PHI [[COPY5]], %bb.2, %8, %bb.3
; CHECK: [[PHI3:%[0-9]+]]:gprlr = PHI [[t2DoLoopStartTP]], %bb.2, %33, %bb.3
; CHECK: [[PHI4:%[0-9]+]]:rgpr = PHI [[COPY6]], %bb.2, %7, %bb.3
; CHECK: [[MVE_VCTP16_:%[0-9]+]]:vccr = MVE_VCTP16 [[PHI4]], 0, $noreg, $noreg
; CHECK: [[t2SUBri1:%[0-9]+]]:rgpr = t2SUBri [[PHI4]], 8, 14 /* CC::al */, $noreg, $noreg
; CHECK: [[COPY7:%[0-9]+]]:gpr = COPY [[t2SUBri1]]
; CHECK: [[MVE_VLDRHU16_post:%[0-9]+]]:rgpr, [[MVE_VLDRHU16_post1:%[0-9]+]]:mqpr = MVE_VLDRHU16_post [[PHI]], 16, 1, [[MVE_VCTP16_]], [[PHI3]] :: (load (s128) from %ir.lsr.iv35, align 2)
; CHECK: [[MVE_VLDRHU16_post2:%[0-9]+]]:rgpr, [[MVE_VLDRHU16_post3:%[0-9]+]]:mqpr = MVE_VLDRHU16_post [[PHI1]], 16, 1, [[MVE_VCTP16_]], [[PHI3]] :: (load (s128) from %ir.lsr.iv12, align 2)
; CHECK: [[MVE_VMLADAVas16_:%[0-9]+]]:tgpreven = MVE_VMLADAVas16 [[PHI2]], killed [[MVE_VLDRHU16_post3]], killed [[MVE_VLDRHU16_post1]], 1, [[MVE_VCTP16_]], [[PHI3]]
; CHECK: [[COPY8:%[0-9]+]]:gpr = COPY [[MVE_VMLADAVas16_]]
; CHECK: [[COPY9:%[0-9]+]]:gpr = COPY [[MVE_VLDRHU16_post2]]
; CHECK: [[COPY10:%[0-9]+]]:gpr = COPY [[MVE_VLDRHU16_post]]
; CHECK: [[t2LoopEndDec:%[0-9]+]]:gprlr = t2LoopEndDec [[PHI3]], %bb.3, implicit-def $cpsr
; CHECK: t2B %bb.4, 14 /* CC::al */, $noreg
; CHECK: bb.4.for.cond.cleanup:
; CHECK: [[PHI5:%[0-9]+]]:gpr = PHI [[COPY3]], %bb.1, [[COPY8]], %bb.3
; CHECK: $r0 = COPY [[PHI5]]
; CHECK: tBX_RET 14 /* CC::al */, $noreg, implicit $r0
bb.0.entry:
successors: %bb.1(0x50000000), %bb.4(0x30000000)
liveins: $r0, $r1, $r2
%15:gprnopc = COPY $r2
%14:gpr = COPY $r1
%13:gpr = COPY $r0
t2CMPri %15, 1, 14 /* CC::al */, $noreg, implicit-def $cpsr
t2Bcc %bb.1, 10 /* CC::ge */, $cpsr
bb.4:
successors: %bb.3(0x80000000)
%22:rgpr = t2MOVi 0, 14 /* CC::al */, $noreg, $noreg
%16:gpr = COPY %22
t2B %bb.3, 14 /* CC::al */, $noreg
bb.1.vector.ph:
successors: %bb.2(0x80000000)
%17:rgpr = t2ADDri %15, 7, 14 /* CC::al */, $noreg, $noreg
%18:rgpr = t2BICri %17, 7, 14 /* CC::al */, $noreg, $noreg
%19:rgpr = t2SUBri %18, 8, 14 /* CC::al */, $noreg, $noreg
%20:rgpr = t2MOVi 1, 14 /* CC::al */, $noreg, $noreg
%21:gprnopc = nuw nsw t2ADDrs %20, %19, 27, 14 /* CC::al */, $noreg, $noreg
%0:rgpr = COPY %21
%24:gprlr = t2DoLoopStart %0
%25:rgpr = t2MOVi 0, 14 /* CC::al */, $noreg, $noreg
%23:gpr = COPY %25
%1:gpr = COPY %24
%35:gprnopc = COPY %15
bb.2.vector.body:
successors: %bb.2(0x7c000000), %bb.3(0x04000000)
%2:rgpr = PHI %13, %bb.1, %10, %bb.2
%3:rgpr = PHI %14, %bb.1, %9, %bb.2
%4:tgpreven = PHI %23, %bb.1, %8, %bb.2
%5:gprlr = PHI %1, %bb.1, %11, %bb.2
%6:rgpr = PHI %35, %bb.1, %7, %bb.2
%26:vccr = MVE_VCTP16 %6, 0, $noreg, $noreg
%27:rgpr = t2SUBri %6, 8, 14 /* CC::al */, $noreg, $noreg
%7:gpr = COPY %27
%28:rgpr, %29:mqpr = MVE_VLDRHU16_post %2, 16, 1, %26, $noreg :: (load (s128) from %ir.lsr.iv35, align 2)
%30:rgpr, %31:mqpr = MVE_VLDRHU16_post %3, 16, 1, %26, $noreg :: (load (s128) from %ir.lsr.iv12, align 2)
%32:tgpreven = MVE_VMLADAVas16 %4, killed %31, killed %29, 1, %26, $noreg
%8:gpr = COPY %32
%9:gpr = COPY %30
%10:gpr = COPY %28
%33:gprlr = t2LoopDec %5, 1
%11:gpr = COPY %33
t2LoopEnd %33, %bb.2, implicit-def dead $cpsr
t2B %bb.3, 14 /* CC::al */, $noreg
bb.3.for.cond.cleanup:
%12:gpr = PHI %16, %bb.4, %8, %bb.2
$r0 = COPY %12
tBX_RET 14 /* CC::al */, $noreg, implicit $r0
...
|
