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
|
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s
;
; LDFF1B, LDFF1W, LDFF1H, LDFF1D: vector base + scalar offset (index)
; e.g. ldff1b { z0.d }, p0/z, [x0, z0.d]
;
; LDFF1B
define <vscale x 4 x i32> @gldff1b_s_scalar_offset(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1b_s_scalar_offset:
; CHECK: ldff1b { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
%res = zext <vscale x 4 x i8> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
define <vscale x 2 x i64> @gldff1b_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1b_d_scalar_offset:
; CHECK: ldff1b { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = zext <vscale x 2 x i8> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
; LDFF1H
define <vscale x 4 x i32> @gldff1h_s_scalar_offset(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1h_s_scalar_offset:
; CHECK: ldff1h { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
%res = zext <vscale x 4 x i16> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
define <vscale x 2 x i64> @gldff1h_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1h_d_scalar_offset:
; CHECK: ldff1h { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = zext <vscale x 2 x i16> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
; LDFF1W
define <vscale x 4 x i32> @gldff1w_s_scalar_offset(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1w_s_scalar_offset:
; CHECK: ldff1w { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
ret <vscale x 4 x i32> %load
}
define <vscale x 2 x i64> @gldff1w_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1w_d_scalar_offset:
; CHECK: ldff1w { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = zext <vscale x 2 x i32> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
define <vscale x 4 x float> @gldff1w_s_scalar_offset_float(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1w_s_scalar_offset_float:
; CHECK: ldff1w { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
ret <vscale x 4 x float> %load
}
; LDFF1D
define <vscale x 2 x i64> @gldff1d_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1d_d_scalar_offset:
; CHECK: ldff1d { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
ret <vscale x 2 x i64> %load
}
define <vscale x 2 x double> @gldff1d_d_scalar_offset_double(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1d_d_scalar_offset_double:
; CHECK: ldff1d { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
ret <vscale x 2 x double> %load
}
; LDFF1SB, LDFF1SW, LDFF1SH: vector base + scalar offset (index)
; e.g. ldff1b { z0.d }, p0/z, [x0, z0.d]
;
; LDFF1SB
define <vscale x 4 x i32> @gldff1sb_s_scalar_offset(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1sb_s_scalar_offset:
; CHECK: ldff1sb { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
%res = sext <vscale x 4 x i8> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
define <vscale x 2 x i64> @gldff1sb_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1sb_d_scalar_offset:
; CHECK: ldff1sb { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = sext <vscale x 2 x i8> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
; LDFF1SH
define <vscale x 4 x i32> @gldff1sh_s_scalar_offset(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %base, i64 %offset) {
; CHECK-LABEL: gldff1sh_s_scalar_offset:
; CHECK: ldff1sh { z0.s }, p0/z, [x0, z0.s, uxtw]
; CHECK-NEXT: ret
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %pg,
<vscale x 4 x i32> %base,
i64 %offset)
%res = sext <vscale x 4 x i16> %load to <vscale x 4 x i32>
ret <vscale x 4 x i32> %res
}
define <vscale x 2 x i64> @gldff1sh_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1sh_d_scalar_offset:
; CHECK: ldff1sh { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = sext <vscale x 2 x i16> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
; LDFF1SW
define <vscale x 2 x i64> @gldff1sw_d_scalar_offset(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %base, i64 %offset) {
; CHECK-LABEL: gldff1sw_d_scalar_offset:
; CHECK: ldff1sw { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT: ret
%load = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 %offset)
%res = sext <vscale x 2 x i32> %load to <vscale x 2 x i64>
ret <vscale x 2 x i64> %res
}
; LDFF1B/LDFF1SB
declare <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, i64)
declare <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
; LDFF1H/LDFF1SH
declare <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, i64)
declare <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
; LDFF1W/LDFF1SW
declare <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, i64)
declare <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
declare <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, i64)
; LDFF1D
declare <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
declare <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
|
