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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve.fp -verify-machineinstrs %s -o - | FileCheck %s
define half @extret1_f16_sf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret1_f16_sf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vmov d0, r0, r1
; CHECK-NEXT: mov r0, sp
; CHECK-NEXT: vldrw.u32 q1, [r0]
; CHECK-NEXT: ldr r0, [sp, #16]
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vmovx.f16 s0, s0
; CHECK-NEXT: vstr.16 s0, [r0]
; CHECK-NEXT: vmov r0, s0
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 1
store half %e, half* %p, align 2
ret half %e
}
define half @extret4_f16_sf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret4_f16_sf:
; CHECK: @ %bb.0:
; CHECK-NEXT: mov r0, sp
; CHECK-NEXT: vmov d1, r2, r3
; CHECK-NEXT: vldrw.u32 q1, [r0]
; CHECK-NEXT: ldr r0, [sp, #16]
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vstr.16 s2, [r0]
; CHECK-NEXT: vmov r0, s2
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 4
store half %e, half* %p, align 2
ret half %e
}
define arm_aapcs_vfpcc half @extret1_f16_hf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret1_f16_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vmovx.f16 s0, s0
; CHECK-NEXT: vstr.16 s0, [r0]
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 1
store half %e, half* %p, align 2
ret half %e
}
define arm_aapcs_vfpcc half @extret4_f16_hf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret4_f16_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vmov.f32 s0, s2
; CHECK-NEXT: vstr.16 s2, [r0]
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 4
store half %e, half* %p, align 2
ret half %e
}
define arm_aapcs_vfpcc <8 x half> @extret1_v8f16_hf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret1_v8f16_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vmov.u16 r1, q0[1]
; CHECK-NEXT: vdup.16 q0, r1
; CHECK-NEXT: strh r1, [r0]
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 1
store half %e, half* %p, align 2
%i = insertelement <8 x half> undef, half %e, i32 0
%s = shufflevector <8 x half> %i, <8 x half> undef, <8 x i32> zeroinitializer
ret <8 x half> %s
}
define arm_aapcs_vfpcc <8 x half> @extret4_v8f16_hf(<8 x half> %a, <8 x half> %b, half* nocapture %p) {
; CHECK-LABEL: extret4_v8f16_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f16 q0, q0, q1
; CHECK-NEXT: vmov.u16 r1, q0[4]
; CHECK-NEXT: vdup.16 q0, r1
; CHECK-NEXT: strh r1, [r0]
; CHECK-NEXT: bx lr
%c = fadd <8 x half> %a, %b
%e = extractelement <8 x half> %c, i32 4
store half %e, half* %p, align 2
%i = insertelement <8 x half> undef, half %e, i32 0
%s = shufflevector <8 x half> %i, <8 x half> undef, <8 x i32> zeroinitializer
ret <8 x half> %s
}
define float @extret1_f32_sf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret1_f32_sf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vmov d0, r0, r1
; CHECK-NEXT: mov r0, sp
; CHECK-NEXT: vldrw.u32 q1, [r0]
; CHECK-NEXT: ldr r1, [sp, #16]
; CHECK-NEXT: vadd.f32 q0, q0, q1
; CHECK-NEXT: vmov r0, s1
; CHECK-NEXT: vstr s1, [r1]
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 1
store float %e, float* %p, align 4
ret float %e
}
define float @extret2_f32_sf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret2_f32_sf:
; CHECK: @ %bb.0:
; CHECK-NEXT: mov r0, sp
; CHECK-NEXT: vmov d1, r2, r3
; CHECK-NEXT: vldrw.u32 q1, [r0]
; CHECK-NEXT: ldr r1, [sp, #16]
; CHECK-NEXT: vadd.f32 q0, q0, q1
; CHECK-NEXT: vmov r0, s2
; CHECK-NEXT: vstr s2, [r1]
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 2
store float %e, float* %p, align 4
ret float %e
}
define arm_aapcs_vfpcc float @extret1_f32_hf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret1_f32_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f32 q0, q0, q1
; CHECK-NEXT: vmov.f32 s0, s1
; CHECK-NEXT: vstr s1, [r0]
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 1
store float %e, float* %p, align 4
ret float %e
}
define arm_aapcs_vfpcc float @extret2_f32_hf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret2_f32_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f32 q0, q0, q1
; CHECK-NEXT: vmov.f32 s0, s2
; CHECK-NEXT: vstr s2, [r0]
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 2
store float %e, float* %p, align 4
ret float %e
}
define arm_aapcs_vfpcc <4 x float> @extret1_v4f32_hf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret1_v4f32_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f32 q1, q0, q1
; CHECK-NEXT: vmov r1, s5
; CHECK-NEXT: vstr s5, [r0]
; CHECK-NEXT: vdup.32 q0, r1
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 1
store float %e, float* %p, align 4
%i = insertelement <4 x float> undef, float %e, i32 0
%s = shufflevector <4 x float> %i, <4 x float> undef, <4 x i32> zeroinitializer
ret <4 x float> %s
}
define arm_aapcs_vfpcc <4 x float> @extret2_v4f32_hf(<4 x float> %a, <4 x float> %b, float* nocapture %p) {
; CHECK-LABEL: extret2_v4f32_hf:
; CHECK: @ %bb.0:
; CHECK-NEXT: vadd.f32 q1, q0, q1
; CHECK-NEXT: vmov r1, s6
; CHECK-NEXT: vstr s6, [r0]
; CHECK-NEXT: vdup.32 q0, r1
; CHECK-NEXT: bx lr
%c = fadd <4 x float> %a, %b
%e = extractelement <4 x float> %c, i32 2
store float %e, float* %p, align 4
%i = insertelement <4 x float> undef, float %e, i32 0
%s = shufflevector <4 x float> %i, <4 x float> undef, <4 x i32> zeroinitializer
ret <4 x float> %s
}
|
