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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
; RUN: llc -mtriple=aarch64 < %s | FileCheck %s
define <16 x i8> @lower_trunc_16xi8(i16 %a, i16 %b, i16 %c, i16 %d, i16 %e, i16 %f, i16 %g, i16 %h, i16 %i, i16 %j, i16 %k, i16 %l, i16 %m, i16 %n, i16 %o, i16 %p) {
; CHECK-LABEL: lower_trunc_16xi8:
; CHECK: // %bb.0:
; CHECK-NEXT: fmov s0, w0
; CHECK-NEXT: ldr h1, [sp]
; CHECK-NEXT: add x8, sp, #8
; CHECK-NEXT: ld1 { v1.h }[1], [x8]
; CHECK-NEXT: add x8, sp, #16
; CHECK-NEXT: mov v0.h[1], w1
; CHECK-NEXT: ld1 { v1.h }[2], [x8]
; CHECK-NEXT: add x8, sp, #24
; CHECK-NEXT: mov v0.h[2], w2
; CHECK-NEXT: ld1 { v1.h }[3], [x8]
; CHECK-NEXT: add x8, sp, #32
; CHECK-NEXT: mov v0.h[3], w3
; CHECK-NEXT: ld1 { v1.h }[4], [x8]
; CHECK-NEXT: add x8, sp, #40
; CHECK-NEXT: ld1 { v1.h }[5], [x8]
; CHECK-NEXT: add x8, sp, #48
; CHECK-NEXT: mov v0.h[4], w4
; CHECK-NEXT: ld1 { v1.h }[6], [x8]
; CHECK-NEXT: add x8, sp, #56
; CHECK-NEXT: mov v0.h[5], w5
; CHECK-NEXT: ld1 { v1.h }[7], [x8]
; CHECK-NEXT: mov v0.h[6], w6
; CHECK-NEXT: add v2.8h, v1.8h, v1.8h
; CHECK-NEXT: mov v0.h[7], w7
; CHECK-NEXT: add v3.8h, v0.8h, v0.8h
; CHECK-NEXT: uzp1 v0.16b, v0.16b, v1.16b
; CHECK-NEXT: uzp1 v1.16b, v3.16b, v2.16b
; CHECK-NEXT: eor v0.16b, v0.16b, v1.16b
; CHECK-NEXT: ret
%a1 = insertelement <16 x i16> poison, i16 %a, i16 0
%b1 = insertelement <16 x i16> %a1, i16 %b, i16 1
%c1 = insertelement <16 x i16> %b1, i16 %c, i16 2
%d1 = insertelement <16 x i16> %c1, i16 %d, i16 3
%e1 = insertelement <16 x i16> %d1, i16 %e, i16 4
%f1 = insertelement <16 x i16> %e1, i16 %f, i16 5
%g1 = insertelement <16 x i16> %f1, i16 %g, i16 6
%h1 = insertelement <16 x i16> %g1, i16 %h, i16 7
%i1 = insertelement <16 x i16> %h1, i16 %i, i16 8
%j1 = insertelement <16 x i16> %i1, i16 %j, i16 9
%k1 = insertelement <16 x i16> %j1, i16 %k, i16 10
%l1 = insertelement <16 x i16> %k1, i16 %l, i16 11
%m1 = insertelement <16 x i16> %l1, i16 %m, i16 12
%n1 = insertelement <16 x i16> %m1, i16 %n, i16 13
%o1 = insertelement <16 x i16> %n1, i16 %o, i16 14
%p1 = insertelement <16 x i16> %o1, i16 %p, i16 15
%t = trunc <16 x i16> %p1 to <16 x i8>
%s = add <16 x i16> %p1, %p1
%t2 = trunc <16 x i16> %s to <16 x i8>
%pro = xor <16 x i8> %t, %t2
ret <16 x i8> %pro
}
define <8 x i16> @lower_trunc_8xi16(i32 %a, i32 %b, i32 %c, i32 %d, i32 %e, i32 %f, i32 %g, i32 %h) {
; CHECK-LABEL: lower_trunc_8xi16:
; CHECK: // %bb.0:
; CHECK-NEXT: fmov s0, w4
; CHECK-NEXT: fmov s1, w0
; CHECK-NEXT: mov v0.s[1], w5
; CHECK-NEXT: mov v1.s[1], w1
; CHECK-NEXT: mov v0.s[2], w6
; CHECK-NEXT: mov v1.s[2], w2
; CHECK-NEXT: mov v0.s[3], w7
; CHECK-NEXT: mov v1.s[3], w3
; CHECK-NEXT: add v2.4s, v0.4s, v0.4s
; CHECK-NEXT: add v3.4s, v1.4s, v1.4s
; CHECK-NEXT: uzp1 v0.8h, v1.8h, v0.8h
; CHECK-NEXT: uzp1 v1.8h, v3.8h, v2.8h
; CHECK-NEXT: eor v0.16b, v0.16b, v1.16b
; CHECK-NEXT: ret
%a1 = insertelement <8 x i32> poison, i32 %a, i32 0
%b1 = insertelement <8 x i32> %a1, i32 %b, i32 1
%c1 = insertelement <8 x i32> %b1, i32 %c, i32 2
%d1 = insertelement <8 x i32> %c1, i32 %d, i32 3
%e1 = insertelement <8 x i32> %d1, i32 %e, i32 4
%f1 = insertelement <8 x i32> %e1, i32 %f, i32 5
%g1 = insertelement <8 x i32> %f1, i32 %g, i32 6
%h1 = insertelement <8 x i32> %g1, i32 %h, i32 7
%t = trunc <8 x i32> %h1 to <8 x i16>
%s = add <8 x i32> %h1, %h1
%t2 = trunc <8 x i32> %s to <8 x i16>
%o = xor <8 x i16> %t, %t2
ret <8 x i16> %o
}
define <4 x i32> @lower_trunc_4xi32(i64 %a, i64 %b, i64 %c, i64 %d) {
; CHECK-LABEL: lower_trunc_4xi32:
; CHECK: // %bb.0:
; CHECK-NEXT: fmov d0, x2
; CHECK-NEXT: fmov d1, x0
; CHECK-NEXT: mov v0.d[1], x3
; CHECK-NEXT: mov v1.d[1], x1
; CHECK-NEXT: add v2.2d, v0.2d, v0.2d
; CHECK-NEXT: add v3.2d, v1.2d, v1.2d
; CHECK-NEXT: uzp1 v0.4s, v1.4s, v0.4s
; CHECK-NEXT: uzp1 v1.4s, v3.4s, v2.4s
; CHECK-NEXT: eor v0.16b, v0.16b, v1.16b
; CHECK-NEXT: ret
%a1 = insertelement <4 x i64> poison, i64 %a, i64 0
%b1 = insertelement <4 x i64> %a1, i64 %b, i64 1
%c1 = insertelement <4 x i64> %b1, i64 %c, i64 2
%d1 = insertelement <4 x i64> %c1, i64 %d, i64 3
%t = trunc <4 x i64> %d1 to <4 x i32>
%s = add <4 x i64> %d1, %d1
%t2 = trunc <4 x i64> %s to <4 x i32>
%o = xor <4 x i32> %t, %t2
ret <4 x i32> %o
}
define <8 x i32> @lower_trunc_8xi32(i64 %a, i64 %b, i64 %c, i64 %d, i64 %e, i64 %f, i64 %g, i64 %h) {
; CHECK-LABEL: lower_trunc_8xi32:
; CHECK: // %bb.0:
; CHECK-NEXT: fmov d0, x2
; CHECK-NEXT: fmov d1, x0
; CHECK-NEXT: fmov d2, x6
; CHECK-NEXT: fmov d3, x4
; CHECK-NEXT: mov v0.d[1], x3
; CHECK-NEXT: mov v1.d[1], x1
; CHECK-NEXT: mov v2.d[1], x7
; CHECK-NEXT: mov v3.d[1], x5
; CHECK-NEXT: add v4.2d, v0.2d, v0.2d
; CHECK-NEXT: add v5.2d, v1.2d, v1.2d
; CHECK-NEXT: add v6.2d, v2.2d, v2.2d
; CHECK-NEXT: add v7.2d, v3.2d, v3.2d
; CHECK-NEXT: uzp1 v2.4s, v3.4s, v2.4s
; CHECK-NEXT: uzp1 v0.4s, v1.4s, v0.4s
; CHECK-NEXT: uzp1 v3.4s, v5.4s, v4.4s
; CHECK-NEXT: uzp1 v1.4s, v7.4s, v6.4s
; CHECK-NEXT: eor v0.16b, v0.16b, v3.16b
; CHECK-NEXT: eor v1.16b, v2.16b, v1.16b
; CHECK-NEXT: ret
%a1 = insertelement <8 x i64> poison, i64 %a, i64 0
%b1 = insertelement <8 x i64> %a1, i64 %b, i64 1
%c1 = insertelement <8 x i64> %b1, i64 %c, i64 2
%d1 = insertelement <8 x i64> %c1, i64 %d, i64 3
%e1 = insertelement <8 x i64> %d1, i64 %e, i64 4
%f1 = insertelement <8 x i64> %e1, i64 %f, i64 5
%g1 = insertelement <8 x i64> %f1, i64 %g, i64 6
%h1 = insertelement <8 x i64> %g1, i64 %h, i64 7
%t = trunc <8 x i64> %h1 to <8 x i32>
%s = add <8 x i64> %h1, %h1
%t2 = trunc <8 x i64> %s to <8 x i32>
%o = xor <8 x i32> %t, %t2
ret <8 x i32> %o
}
|
