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
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s
; Verify that DAG combine rules for LD1 + sext/zext don't apply when the
; result of LD1 has multiple uses
define <vscale x 2 x i64> @no_dag_combine_zext_sext(<vscale x 2 x i1> %pg,
; CHECK-LABEL: no_dag_combine_zext_sext:
; CHECK: // %bb.0:
; CHECK-NEXT: ld1b { z0.d }, p0/z, [z0.d, #16]
; CHECK-NEXT: st1b { z0.d }, p1, [x0]
; CHECK-NEXT: and z0.d, z0.d, #0xff
; CHECK-NEXT: ret
<vscale x 2 x i64> %base,
ptr %res_out,
<vscale x 2 x i1> %pred) {
%load = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 16)
%res1 = zext <vscale x 2 x i8> %load to <vscale x 2 x i64>
%res2 = sext <vscale x 2 x i8> %load to <vscale x 2 x i64>
call void @llvm.masked.store.nxv2i8(<vscale x 2 x i8> %load,
ptr %res_out,
i32 8,
<vscale x 2 x i1> %pred)
ret <vscale x 2 x i64> %res1
}
define <vscale x 2 x i64> @no_dag_combine_sext(<vscale x 2 x i1> %pg,
; CHECK-LABEL: no_dag_combine_sext:
; CHECK: // %bb.0:
; CHECK-NEXT: ld1b { z1.d }, p0/z, [z0.d, #16]
; CHECK-NEXT: ptrue p0.d
; CHECK-NEXT: movprfx z0, z1
; CHECK-NEXT: sxtb z0.d, p0/m, z1.d
; CHECK-NEXT: st1b { z1.d }, p1, [x0]
; CHECK-NEXT: ret
<vscale x 2 x i64> %base,
ptr %res_out,
<vscale x 2 x i1> %pred) {
%load = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 16)
%res = sext <vscale x 2 x i8> %load to <vscale x 2 x i64>
call void @llvm.masked.store.nxv2i8(<vscale x 2 x i8> %load,
ptr %res_out,
i32 8,
<vscale x 2 x i1> %pred)
ret <vscale x 2 x i64> %res
}
define <vscale x 2 x i64> @no_dag_combine_zext(<vscale x 2 x i1> %pg,
; CHECK-LABEL: no_dag_combine_zext:
; CHECK: // %bb.0:
; CHECK-NEXT: ld1b { z0.d }, p0/z, [z0.d, #16]
; CHECK-NEXT: st1b { z0.d }, p1, [x0]
; CHECK-NEXT: and z0.d, z0.d, #0xff
; CHECK-NEXT: ret
<vscale x 2 x i64> %base,
ptr %res_out,
<vscale x 2 x i1> %pred) {
%load = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %pg,
<vscale x 2 x i64> %base,
i64 16)
%res = zext <vscale x 2 x i8> %load to <vscale x 2 x i64>
call void @llvm.masked.store.nxv2i8(<vscale x 2 x i8> %load,
ptr %res_out,
i32 8,
<vscale x 2 x i1> %pred)
ret <vscale x 2 x i64> %res
}
define <vscale x 16 x i8> @narrow_i64_gather_index_i8_zext(ptr %out, ptr %in, <vscale x 16 x i8> %d, i64 %ptr){
; CHECK-LABEL: narrow_i64_gather_index_i8_zext:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: add x8, x1, x2
; CHECK-NEXT: ld1b { z0.s }, p0/z, [x8, #3, mul vl]
; CHECK-NEXT: ld1b { z1.s }, p0/z, [x8, #2, mul vl]
; CHECK-NEXT: ld1b { z2.s }, p0/z, [x1, x2]
; CHECK-NEXT: ld1b { z3.s }, p0/z, [x8, #1, mul vl]
; CHECK-NEXT: ld1b { z0.s }, p0/z, [x1, z0.s, uxtw]
; CHECK-NEXT: ld1b { z1.s }, p0/z, [x1, z1.s, uxtw]
; CHECK-NEXT: ld1b { z2.s }, p0/z, [x1, z2.s, uxtw]
; CHECK-NEXT: ld1b { z3.s }, p0/z, [x1, z3.s, uxtw]
; CHECK-NEXT: uzp1 z0.h, z1.h, z0.h
; CHECK-NEXT: uzp1 z1.h, z2.h, z3.h
; CHECK-NEXT: uzp1 z0.b, z1.b, z0.b
; CHECK-NEXT: ret
%1 = getelementptr inbounds i8, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 16 x i8>, ptr %2, align 1
%3 = zext <vscale x 16 x i8> %wide.load to <vscale x 16 x i64>
%4 = getelementptr inbounds i8, ptr %in, <vscale x 16 x i64> %3
%wide.masked.gather = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> %4, i32 1, <vscale x 16 x i1> splat (i1 true), <vscale x 16 x i8> poison)
ret <vscale x 16 x i8> %wide.masked.gather
}
define <vscale x 16 x i8> @narrow_i64_gather_index_i8_sext(ptr %out, ptr %in, <vscale x 16 x i8> %d, i64 %ptr){
; CHECK-LABEL: narrow_i64_gather_index_i8_sext:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: add x8, x1, x2
; CHECK-NEXT: ld1sb { z0.s }, p0/z, [x8, #3, mul vl]
; CHECK-NEXT: ld1sb { z1.s }, p0/z, [x8, #2, mul vl]
; CHECK-NEXT: ld1sb { z2.s }, p0/z, [x1, x2]
; CHECK-NEXT: ld1sb { z3.s }, p0/z, [x8, #1, mul vl]
; CHECK-NEXT: ld1b { z0.s }, p0/z, [x1, z0.s, sxtw]
; CHECK-NEXT: ld1b { z1.s }, p0/z, [x1, z1.s, sxtw]
; CHECK-NEXT: ld1b { z2.s }, p0/z, [x1, z2.s, sxtw]
; CHECK-NEXT: ld1b { z3.s }, p0/z, [x1, z3.s, sxtw]
; CHECK-NEXT: uzp1 z0.h, z1.h, z0.h
; CHECK-NEXT: uzp1 z1.h, z2.h, z3.h
; CHECK-NEXT: uzp1 z0.b, z1.b, z0.b
; CHECK-NEXT: ret
%1 = getelementptr inbounds i8, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 16 x i8>, ptr %2, align 1
%3 = sext <vscale x 16 x i8> %wide.load to <vscale x 16 x i64>
%4 = getelementptr inbounds i8, ptr %in, <vscale x 16 x i64> %3
%wide.masked.gather = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> %4, i32 1, <vscale x 16 x i1> splat (i1 true), <vscale x 16 x i8> poison)
ret <vscale x 16 x i8> %wide.masked.gather
}
define <vscale x 8 x i16> @narrow_i64_gather_index_i16_zext(ptr %out, ptr %in, <vscale x 8 x i16> %d, i64 %ptr){
; CHECK-LABEL: narrow_i64_gather_index_i16_zext:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: add x8, x1, x2, lsl #1
; CHECK-NEXT: ld1h { z0.s }, p0/z, [x1, x2, lsl #1]
; CHECK-NEXT: ld1h { z1.s }, p0/z, [x8, #1, mul vl]
; CHECK-NEXT: ld1h { z0.s }, p0/z, [x1, z0.s, uxtw #1]
; CHECK-NEXT: ld1h { z1.s }, p0/z, [x1, z1.s, uxtw #1]
; CHECK-NEXT: uzp1 z0.h, z0.h, z1.h
; CHECK-NEXT: ret
%1 = getelementptr inbounds i16, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 8 x i16>, ptr %2, align 1
%3 = zext <vscale x 8 x i16> %wide.load to <vscale x 8 x i64>
%4 = getelementptr inbounds i16, ptr %in, <vscale x 8 x i64> %3
%wide.masked.gather = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> %4, i32 1, <vscale x 8 x i1> splat (i1 true), <vscale x 8 x i16> poison)
ret <vscale x 8 x i16> %wide.masked.gather
}
define <vscale x 8 x i16> @narrow_i64_gather_index_i16_sext(ptr %out, ptr %in, <vscale x 8 x i16> %d, i64 %ptr){
; CHECK-LABEL: narrow_i64_gather_index_i16_sext:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: add x8, x1, x2, lsl #1
; CHECK-NEXT: ld1sh { z0.s }, p0/z, [x1, x2, lsl #1]
; CHECK-NEXT: ld1sh { z1.s }, p0/z, [x8, #1, mul vl]
; CHECK-NEXT: ld1h { z0.s }, p0/z, [x1, z0.s, sxtw #1]
; CHECK-NEXT: ld1h { z1.s }, p0/z, [x1, z1.s, sxtw #1]
; CHECK-NEXT: uzp1 z0.h, z0.h, z1.h
; CHECK-NEXT: ret
%1 = getelementptr inbounds i16, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 8 x i16>, ptr %2, align 1
%3 = sext <vscale x 8 x i16> %wide.load to <vscale x 8 x i64>
%4 = getelementptr inbounds i16, ptr %in, <vscale x 8 x i64> %3
%wide.masked.gather = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> %4, i32 1, <vscale x 8 x i1> splat (i1 true), <vscale x 8 x i16> poison)
ret <vscale x 8 x i16> %wide.masked.gather
}
define <vscale x 4 x i32> @no_narrow_i64_gather_index_i32(ptr %out, ptr %in, <vscale x 4 x i32> %d, i64 %ptr){
; CHECK-LABEL: no_narrow_i64_gather_index_i32:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: ld1w { z0.s }, p0/z, [x1, x2, lsl #2]
; CHECK-NEXT: ld1w { z0.s }, p0/z, [x1, z0.s, uxtw #2]
; CHECK-NEXT: ret
%1 = getelementptr inbounds i32, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 4 x i32>, ptr %2, align 1
%3 = zext <vscale x 4 x i32> %wide.load to <vscale x 4 x i64>
%4 = getelementptr inbounds i32, ptr %in, <vscale x 4 x i64> %3
%wide.masked.gather = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> %4, i32 1, <vscale x 4 x i1> splat (i1 true), <vscale x 4 x i32> poison)
ret <vscale x 4 x i32> %wide.masked.gather
}
define <vscale x 2 x i64> @no_narrow_i64_gather_index_i64(ptr %out, ptr %in, <vscale x 2 x i64> %d, i64 %ptr){
; CHECK-LABEL: no_narrow_i64_gather_index_i64:
; CHECK: // %bb.0:
; CHECK-NEXT: ptrue p0.d
; CHECK-NEXT: ld1d { z0.d }, p0/z, [x1, x2, lsl #3]
; CHECK-NEXT: ld1d { z0.d }, p0/z, [x1, z0.d, lsl #3]
; CHECK-NEXT: ret
%1 = getelementptr inbounds i64, ptr %in, i64 %ptr
%2 = bitcast ptr %1 to ptr
%wide.load = load <vscale x 2 x i64>, ptr %2, align 1
%3 = getelementptr inbounds i64, ptr %in, <vscale x 2 x i64> %wide.load
%wide.masked.gather = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> %3, i32 1, <vscale x 2 x i1> splat (i1 true), <vscale x 2 x i64> poison)
ret <vscale x 2 x i64> %wide.masked.gather
}
declare <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, i64)
declare void @llvm.masked.store.nxv2i8(<vscale x 2 x i8>, ptr, i32, <vscale x 2 x i1>)
declare <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr>, i32, <vscale x 16 x i1>, <vscale x 16 x i8>)
declare <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr>, i32, <vscale x 8 x i1>, <vscale x 8 x i16>)
declare <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr>, i32, <vscale x 4 x i1>, <vscale x 4 x i32>)
declare <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x i64>)
|
