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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+lsx < %s | FileCheck %s
; RUN: llc --mtriple=loongarch64 --mattr=+lsx < %s | FileCheck %s
define void @minnum_v4f32(ptr %res, ptr %x, ptr %y) nounwind {
; CHECK-LABEL: minnum_v4f32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vld $vr0, $a2, 0
; CHECK-NEXT: vld $vr1, $a1, 0
; CHECK-NEXT: vreplvei.w $vr2, $vr0, 1
; CHECK-NEXT: vreplvei.w $vr3, $vr1, 1
; CHECK-NEXT: fmin.s $fa2, $fa3, $fa2
; CHECK-NEXT: vreplvei.w $vr3, $vr0, 0
; CHECK-NEXT: vreplvei.w $vr4, $vr1, 0
; CHECK-NEXT: fmin.s $fa3, $fa4, $fa3
; CHECK-NEXT: vextrins.w $vr3, $vr2, 16
; CHECK-NEXT: vreplvei.w $vr2, $vr0, 2
; CHECK-NEXT: vreplvei.w $vr4, $vr1, 2
; CHECK-NEXT: fmin.s $fa2, $fa4, $fa2
; CHECK-NEXT: vextrins.w $vr3, $vr2, 32
; CHECK-NEXT: vreplvei.w $vr0, $vr0, 3
; CHECK-NEXT: vreplvei.w $vr1, $vr1, 3
; CHECK-NEXT: fmin.s $fa0, $fa1, $fa0
; CHECK-NEXT: vextrins.w $vr3, $vr0, 48
; CHECK-NEXT: vst $vr3, $a0, 0
; CHECK-NEXT: ret
entry:
%v0 = load <4 x float>, ptr %x
%v1 = load <4 x float>, ptr %y
%r = call <4 x float> @llvm.minnum.v4f32(<4 x float> %v0, <4 x float> %v1)
store <4 x float> %r, ptr %res
ret void
}
define void @minnum_v2f64(ptr %res, ptr %x, ptr %y) nounwind {
; CHECK-LABEL: minnum_v2f64:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vld $vr0, $a2, 0
; CHECK-NEXT: vld $vr1, $a1, 0
; CHECK-NEXT: vreplvei.d $vr2, $vr0, 1
; CHECK-NEXT: vreplvei.d $vr3, $vr1, 1
; CHECK-NEXT: fmin.d $fa2, $fa3, $fa2
; CHECK-NEXT: vreplvei.d $vr0, $vr0, 0
; CHECK-NEXT: vreplvei.d $vr1, $vr1, 0
; CHECK-NEXT: fmin.d $fa0, $fa1, $fa0
; CHECK-NEXT: vextrins.d $vr0, $vr2, 16
; CHECK-NEXT: vst $vr0, $a0, 0
; CHECK-NEXT: ret
entry:
%v0 = load <2 x double>, ptr %x
%v1 = load <2 x double>, ptr %y
%r = call <2 x double> @llvm.minnum.v2f64(<2 x double> %v0, <2 x double> %v1)
store <2 x double> %r, ptr %res
ret void
}
define void @maxnum_v4f32(ptr %res, ptr %x, ptr %y) nounwind {
; CHECK-LABEL: maxnum_v4f32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vld $vr0, $a2, 0
; CHECK-NEXT: vld $vr1, $a1, 0
; CHECK-NEXT: vreplvei.w $vr2, $vr0, 1
; CHECK-NEXT: vreplvei.w $vr3, $vr1, 1
; CHECK-NEXT: fmax.s $fa2, $fa3, $fa2
; CHECK-NEXT: vreplvei.w $vr3, $vr0, 0
; CHECK-NEXT: vreplvei.w $vr4, $vr1, 0
; CHECK-NEXT: fmax.s $fa3, $fa4, $fa3
; CHECK-NEXT: vextrins.w $vr3, $vr2, 16
; CHECK-NEXT: vreplvei.w $vr2, $vr0, 2
; CHECK-NEXT: vreplvei.w $vr4, $vr1, 2
; CHECK-NEXT: fmax.s $fa2, $fa4, $fa2
; CHECK-NEXT: vextrins.w $vr3, $vr2, 32
; CHECK-NEXT: vreplvei.w $vr0, $vr0, 3
; CHECK-NEXT: vreplvei.w $vr1, $vr1, 3
; CHECK-NEXT: fmax.s $fa0, $fa1, $fa0
; CHECK-NEXT: vextrins.w $vr3, $vr0, 48
; CHECK-NEXT: vst $vr3, $a0, 0
; CHECK-NEXT: ret
entry:
%v0 = load <4 x float>, ptr %x
%v1 = load <4 x float>, ptr %y
%r = call <4 x float> @llvm.maxnum.v4f32(<4 x float> %v0, <4 x float> %v1)
store <4 x float> %r, ptr %res
ret void
}
define void @maxnum_v2f64(ptr %res, ptr %x, ptr %y) nounwind {
; CHECK-LABEL: maxnum_v2f64:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vld $vr0, $a2, 0
; CHECK-NEXT: vld $vr1, $a1, 0
; CHECK-NEXT: vreplvei.d $vr2, $vr0, 1
; CHECK-NEXT: vreplvei.d $vr3, $vr1, 1
; CHECK-NEXT: fmax.d $fa2, $fa3, $fa2
; CHECK-NEXT: vreplvei.d $vr0, $vr0, 0
; CHECK-NEXT: vreplvei.d $vr1, $vr1, 0
; CHECK-NEXT: fmax.d $fa0, $fa1, $fa0
; CHECK-NEXT: vextrins.d $vr0, $vr2, 16
; CHECK-NEXT: vst $vr0, $a0, 0
; CHECK-NEXT: ret
entry:
%v0 = load <2 x double>, ptr %x
%v1 = load <2 x double>, ptr %y
%r = call <2 x double> @llvm.maxnum.v2f64(<2 x double> %v0, <2 x double> %v1)
store <2 x double> %r, ptr %res
ret void
}
declare <4 x float> @llvm.minnum.v4f32(<4 x float>, <4 x float>)
declare <2 x double> @llvm.minnum.v2f64(<2 x double>, <2 x double>)
declare <4 x float> @llvm.maxnum.v4f32(<4 x float>, <4 x float>)
declare <2 x double> @llvm.maxnum.v2f64(<2 x double>, <2 x double>)
|
