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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=aarch64 | FileCheck %s --check-prefixes=CHECK,CHECK-SD
; RUN: llc < %s -mtriple=aarch64 -global-isel | FileCheck %s --check-prefixes=CHECK,CHECK-GI
define i64 @select_ogt_float(float %a, float %b) {
; CHECK-SD-LABEL: select_ogt_float:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: fcmp s0, s1
; CHECK-SD-NEXT: mov w8, #4 // =0x4
; CHECK-SD-NEXT: csel x0, x8, xzr, gt
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_ogt_float:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: fcmp s0, s1
; CHECK-GI-NEXT: cset w8, gt
; CHECK-GI-NEXT: lsl x0, x8, #2
; CHECK-GI-NEXT: ret
entry:
%cc = fcmp ogt float %a, %b
%sel = select i1 %cc, i64 4, i64 0
ret i64 %sel
}
define i64 @select_ule_float_inverse(float %a, float %b) {
; CHECK-SD-LABEL: select_ule_float_inverse:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: fcmp s0, s1
; CHECK-SD-NEXT: mov w8, #4 // =0x4
; CHECK-SD-NEXT: csel x0, xzr, x8, le
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_ule_float_inverse:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: fcmp s0, s1
; CHECK-GI-NEXT: cset w8, gt
; CHECK-GI-NEXT: lsl x0, x8, #2
; CHECK-GI-NEXT: ret
entry:
%cc = fcmp ule float %a, %b
%sel = select i1 %cc, i64 0, i64 4
ret i64 %sel
}
define i64 @select_eq_i32(i32 %a, i32 %b) {
; CHECK-SD-LABEL: select_eq_i32:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: mov w8, #4 // =0x4
; CHECK-SD-NEXT: cmp w0, w1
; CHECK-SD-NEXT: csel x0, x8, xzr, eq
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_eq_i32:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: cmp w0, w1
; CHECK-GI-NEXT: cset w8, eq
; CHECK-GI-NEXT: lsl x0, x8, #2
; CHECK-GI-NEXT: ret
entry:
%cc = icmp eq i32 %a, %b
%sel = select i1 %cc, i64 4, i64 0
ret i64 %sel
}
define i64 @select_ne_i32_inverse(i32 %a, i32 %b) {
; CHECK-SD-LABEL: select_ne_i32_inverse:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: mov w8, #4 // =0x4
; CHECK-SD-NEXT: cmp w0, w1
; CHECK-SD-NEXT: csel x0, xzr, x8, ne
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_ne_i32_inverse:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: cmp w0, w1
; CHECK-GI-NEXT: cset w8, eq
; CHECK-GI-NEXT: lsl x0, x8, #2
; CHECK-GI-NEXT: ret
entry:
%cc = icmp ne i32 %a, %b
%sel = select i1 %cc, i64 0, i64 4
ret i64 %sel
}
define <2 x double> @select_olt_load_cmp(<2 x double> %a, ptr %src) {
; CHECK-SD-LABEL: select_olt_load_cmp:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: ldr d1, [x0]
; CHECK-SD-NEXT: fcmgt v1.2s, v1.2s, #0.0
; CHECK-SD-NEXT: sshll v1.2d, v1.2s, #0
; CHECK-SD-NEXT: and v0.16b, v1.16b, v0.16b
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_olt_load_cmp:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: ldr d1, [x0]
; CHECK-GI-NEXT: movi v2.2d, #0000000000000000
; CHECK-GI-NEXT: fcmgt v1.2s, v1.2s, #0.0
; CHECK-GI-NEXT: ushll v1.2d, v1.2s, #0
; CHECK-GI-NEXT: shl v1.2d, v1.2d, #63
; CHECK-GI-NEXT: sshr v1.2d, v1.2d, #63
; CHECK-GI-NEXT: bif v0.16b, v2.16b, v1.16b
; CHECK-GI-NEXT: ret
entry:
%l = load <2 x float>, ptr %src, align 4
%cmp = fcmp olt <2 x float> zeroinitializer, %l
%sel = select <2 x i1> %cmp, <2 x double> %a, <2 x double> zeroinitializer
ret <2 x double> %sel
}
define <4 x i32> @select_icmp_sgt(<4 x i32> %a, <4 x i8> %b) {
; CHECK-SD-LABEL: select_icmp_sgt:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: shl v1.4h, v1.4h, #8
; CHECK-SD-NEXT: sshr v1.4h, v1.4h, #8
; CHECK-SD-NEXT: cmgt v1.4h, v1.4h, #0
; CHECK-SD-NEXT: sshll v1.4s, v1.4h, #0
; CHECK-SD-NEXT: bic v0.16b, v0.16b, v1.16b
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: select_icmp_sgt:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: mov w8, #0 // =0x0
; CHECK-GI-NEXT: uzp1 v1.8b, v1.8b, v0.8b
; CHECK-GI-NEXT: fmov s2, w8
; CHECK-GI-NEXT: mov v2.b[1], w8
; CHECK-GI-NEXT: mov v2.b[2], w8
; CHECK-GI-NEXT: mov v2.b[3], w8
; CHECK-GI-NEXT: cmgt v1.8b, v1.8b, v2.8b
; CHECK-GI-NEXT: umov w8, v1.b[0]
; CHECK-GI-NEXT: umov w9, v1.b[1]
; CHECK-GI-NEXT: fmov s2, w8
; CHECK-GI-NEXT: umov w8, v1.b[2]
; CHECK-GI-NEXT: mov v2.s[1], w9
; CHECK-GI-NEXT: umov w9, v1.b[3]
; CHECK-GI-NEXT: mov v2.s[2], w8
; CHECK-GI-NEXT: mov v2.s[3], w9
; CHECK-GI-NEXT: shl v1.4s, v2.4s, #31
; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31
; CHECK-GI-NEXT: bic v0.16b, v0.16b, v1.16b
; CHECK-GI-NEXT: ret
entry:
%cmp = icmp sgt <4 x i8> %b, zeroinitializer
%sel = select <4 x i1> %cmp, <4 x i32> zeroinitializer, <4 x i32> %a
ret <4 x i32> %sel
}
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; CHECK: {{.*}}
|
