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
|
; RUN: opt -S -instcombine < %s | FileCheck %s
define <4 x i32> @mulByZero(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> %x, <4 x i16> zeroinitializer) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> zeroinitializer
}
define <4 x i32> @mulByOne(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> %x, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: %a = sext <4 x i16> %x to <4 x i32>
; CHECK-NEXT: ret <4 x i32> %a
}
define <4 x i32> @constantMul() nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> <i16 3, i16 3, i16 3, i16 3>, <4 x i16> <i16 2, i16 2, i16 2, i16 2>) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 6, i32 6, i32 6, i32 6>
}
define <4 x i32> @constantMulS() nounwind readnone ssp {
entry:
%b = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> <i16 -1, i16 -1, i16 -1, i16 -1>, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>
}
define <4 x i32> @constantMulU() nounwind readnone ssp {
entry:
%b = tail call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> <i16 -1, i16 -1, i16 -1, i16 -1>, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535>
}
define <4 x i32> @complex1(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> <i16 2, i16 2, i16 2, i16 2>, <4 x i16> %x) nounwind
%b = add <4 x i32> zeroinitializer, %a
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: %a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> <i16 2, i16 2, i16 2, i16 2>, <4 x i16> %x) [[NUW:#[0-9]+]]
; CHECK-NEXT: ret <4 x i32> %a
}
define <4 x i32> @complex2(<4 x i32> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> <i16 3, i16 3, i16 3, i16 3>, <4 x i16> <i16 2, i16 2, i16 2, i16 2>) nounwind
%b = add <4 x i32> %x, %a
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: %b = add <4 x i32> %x, <i32 6, i32 6, i32 6, i32 6>
; CHECK-NEXT: ret <4 x i32> %b
}
declare <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
; ARM64 variants - <rdar://problem/12349617>
define <4 x i32> @mulByZeroARM64(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> %x, <4 x i16> zeroinitializer) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> zeroinitializer
}
define <4 x i32> @mulByOneARM64(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> %x, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: %a = sext <4 x i16> %x to <4 x i32>
; CHECK-NEXT: ret <4 x i32> %a
}
define <4 x i32> @constantMulARM64() nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> <i16 3, i16 3, i16 3, i16 3>, <4 x i16> <i16 2, i16 2, i16 2, i16 2>) nounwind
ret <4 x i32> %a
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 6, i32 6, i32 6, i32 6>
}
define <4 x i32> @constantMulSARM64() nounwind readnone ssp {
entry:
%b = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> <i16 -1, i16 -1, i16 -1, i16 -1>, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>
}
define <4 x i32> @constantMulUARM64() nounwind readnone ssp {
entry:
%b = tail call <4 x i32> @llvm.aarch64.neon.umull.v4i32(<4 x i16> <i16 -1, i16 -1, i16 -1, i16 -1>, <4 x i16> <i16 1, i16 1, i16 1, i16 1>) nounwind
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: ret <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535>
}
define <4 x i32> @complex1ARM64(<4 x i16> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> <i16 2, i16 2, i16 2, i16 2>, <4 x i16> %x) nounwind
%b = add <4 x i32> zeroinitializer, %a
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: %a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> <i16 2, i16 2, i16 2, i16 2>, <4 x i16> %x) [[NUW:#[0-9]+]]
; CHECK-NEXT: ret <4 x i32> %a
}
define <4 x i32> @complex2ARM64(<4 x i32> %x) nounwind readnone ssp {
entry:
%a = tail call <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16> <i16 3, i16 3, i16 3, i16 3>, <4 x i16> <i16 2, i16 2, i16 2, i16 2>) nounwind
%b = add <4 x i32> %x, %a
ret <4 x i32> %b
; CHECK: entry:
; CHECK-NEXT: %b = add <4 x i32> %x, <i32 6, i32 6, i32 6, i32 6>
; CHECK-NEXT: ret <4 x i32> %b
}
declare <4 x i32> @llvm.aarch64.neon.smull.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
declare <4 x i32> @llvm.aarch64.neon.umull.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
; CHECK: attributes #0 = { nounwind readnone ssp }
; CHECK: attributes #1 = { nounwind readnone }
; CHECK: attributes [[NUW]] = { nounwind }
|
