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
|
; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
; This tests fcmp operations that do not map directly to NEON instructions.
; une is implemented with VCEQ/VMVN
define <2 x i32> @vcunef32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcunef32:
;CHECK: vceq.f32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp une <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; olt is implemented with VCGT
define <2 x i32> @vcoltf32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcoltf32:
;CHECK: vcgt.f32
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp olt <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ole is implemented with VCGE
define <2 x i32> @vcolef32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcolef32:
;CHECK: vcge.f32
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ole <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; uge is implemented with VCGT/VMVN
define <2 x i32> @vcugef32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcugef32:
;CHECK: vcgt.f32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp uge <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ule is implemented with VCGT/VMVN
define <2 x i32> @vculef32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vculef32:
;CHECK: vcgt.f32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ule <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ugt is implemented with VCGE/VMVN
define <2 x i32> @vcugtf32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcugtf32:
;CHECK: vcge.f32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ugt <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ult is implemented with VCGE/VMVN
define <2 x i32> @vcultf32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcultf32:
;CHECK: vcge.f32
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ult <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ueq is implemented with VCGT/VCGT/VORR/VMVN
define <2 x i32> @vcueqf32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcueqf32:
;CHECK: vcgt.f32
;CHECK-NEXT: vcgt.f32
;CHECK-NEXT: vorr
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ueq <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; one is implemented with VCGT/VCGT/VORR
define <2 x i32> @vconef32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vconef32:
;CHECK: vcgt.f32
;CHECK-NEXT: vcgt.f32
;CHECK-NEXT: vorr
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp one <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; uno is implemented with VCGT/VCGE/VORR/VMVN
define <2 x i32> @vcunof32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcunof32:
;CHECK: vcge.f32
;CHECK-NEXT: vcgt.f32
;CHECK-NEXT: vorr
;CHECK-NEXT: vmvn
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp uno <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
; ord is implemented with VCGT/VCGE/VORR
define <2 x i32> @vcordf32(<2 x float>* %A, <2 x float>* %B) nounwind {
;CHECK-LABEL: vcordf32:
;CHECK: vcge.f32
;CHECK-NEXT: vcgt.f32
;CHECK-NEXT: vorr
%tmp1 = load <2 x float>, <2 x float>* %A
%tmp2 = load <2 x float>, <2 x float>* %B
%tmp3 = fcmp ord <2 x float> %tmp1, %tmp2
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
ret <2 x i32> %tmp4
}
|
