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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
|
; RUN: llc < %s -march=r600 -mcpu=SI | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc < %s -march=r600 -mcpu=cypress | FileCheck -check-prefix=EG -check-prefix=FUNC %s
declare i32 @llvm.AMDGPU.imax(i32, i32) nounwind readnone
; FUNC-LABEL: @sext_in_reg_i1_i32
; SI: S_LOAD_DWORD [[ARG:s[0-9]+]],
; SI: V_BFE_I32 [[EXTRACT:v[0-9]+]], [[ARG]], 0, 1
; SI: BUFFER_STORE_DWORD [[EXTRACT]],
; EG: BFE_INT
define void @sext_in_reg_i1_i32(i32 addrspace(1)* %out, i32 %in) {
%shl = shl i32 %in, 31
%sext = ashr i32 %shl, 31
store i32 %sext, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: @sext_in_reg_i8_to_i32
; SI: S_ADD_I32 [[VAL:s[0-9]+]],
; SI: V_BFE_I32 [[EXTRACT:v[0-9]+]], [[VAL]], 0, 8
; SI: BUFFER_STORE_DWORD [[EXTRACT]],
; EG: BFE_INT
define void @sext_in_reg_i8_to_i32(i32 addrspace(1)* %out, i32 %a, i32 %b) nounwind {
%c = add i32 %a, %b ; add to prevent folding into extload
%shl = shl i32 %c, 24
%ashr = ashr i32 %shl, 24
store i32 %ashr, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @sext_in_reg_i16_to_i32
; SI: S_ADD_I32 [[VAL:s[0-9]+]],
; SI: V_BFE_I32 [[EXTRACT:v[0-9]+]], [[VAL]], 0, 16
; SI: BUFFER_STORE_DWORD [[EXTRACT]],
; EG: BFE_INT
define void @sext_in_reg_i16_to_i32(i32 addrspace(1)* %out, i32 %a, i32 %b) nounwind {
%c = add i32 %a, %b ; add to prevent folding into extload
%shl = shl i32 %c, 16
%ashr = ashr i32 %shl, 16
store i32 %ashr, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @sext_in_reg_i8_to_v1i32
; SI: S_ADD_I32 [[VAL:s[0-9]+]],
; SI: V_BFE_I32 [[EXTRACT:v[0-9]+]], [[VAL]], 0, 8
; SI: BUFFER_STORE_DWORD [[EXTRACT]],
; EG: BFE_INT
define void @sext_in_reg_i8_to_v1i32(<1 x i32> addrspace(1)* %out, <1 x i32> %a, <1 x i32> %b) nounwind {
%c = add <1 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <1 x i32> %c, <i32 24>
%ashr = ashr <1 x i32> %shl, <i32 24>
store <1 x i32> %ashr, <1 x i32> addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @sext_in_reg_i8_to_i64
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_ASHRREV_I32_e32 {{v[0-9]+}}, 31,
; SI: BUFFER_STORE_DWORD
; EG: BFE_INT
; EG: ASHR
define void @sext_in_reg_i8_to_i64(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
%c = add i64 %a, %b
%shl = shl i64 %c, 56
%ashr = ashr i64 %shl, 56
store i64 %ashr, i64 addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_i16_to_i64
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 16
; SI: V_ASHRREV_I32_e32 {{v[0-9]+}}, 31,
; SI: BUFFER_STORE_DWORD
; EG: BFE_INT
; EG: ASHR
define void @sext_in_reg_i16_to_i64(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
%c = add i64 %a, %b
%shl = shl i64 %c, 48
%ashr = ashr i64 %shl, 48
store i64 %ashr, i64 addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_i32_to_i64
; SI: S_LOAD_DWORD
; SI: S_LOAD_DWORD
; SI: S_ADD_I32 [[ADD:s[0-9]+]],
; SI: S_ASHR_I32 s{{[0-9]+}}, [[ADD]], 31
; SI: BUFFER_STORE_DWORDX2
define void @sext_in_reg_i32_to_i64(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
%c = add i64 %a, %b
%shl = shl i64 %c, 32
%ashr = ashr i64 %shl, 32
store i64 %ashr, i64 addrspace(1)* %out, align 8
ret void
}
; This is broken on Evergreen for some reason related to the <1 x i64> kernel arguments.
; XFUNC-LABEL: @sext_in_reg_i8_to_v1i64
; XSI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; XSI: V_ASHRREV_I32_e32 {{v[0-9]+}}, 31,
; XSI: BUFFER_STORE_DWORD
; XEG: BFE_INT
; XEG: ASHR
; define void @sext_in_reg_i8_to_v1i64(<1 x i64> addrspace(1)* %out, <1 x i64> %a, <1 x i64> %b) nounwind {
; %c = add <1 x i64> %a, %b
; %shl = shl <1 x i64> %c, <i64 56>
; %ashr = ashr <1 x i64> %shl, <i64 56>
; store <1 x i64> %ashr, <1 x i64> addrspace(1)* %out, align 8
; ret void
; }
; FUNC-LABEL: @sext_in_reg_i1_in_i32_other_amount
; SI-NOT: BFE
; SI: S_LSHL_B32 [[REG:s[0-9]+]], {{s[0-9]+}}, 6
; SI: S_ASHR_I32 {{s[0-9]+}}, [[REG]], 7
; EG-NOT: BFE
define void @sext_in_reg_i1_in_i32_other_amount(i32 addrspace(1)* %out, i32 %a, i32 %b) nounwind {
%c = add i32 %a, %b
%x = shl i32 %c, 6
%y = ashr i32 %x, 7
store i32 %y, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: @sext_in_reg_v2i1_in_v2i32_other_amount
; SI: S_LSHL_B32 [[REG0:s[0-9]+]], {{s[0-9]}}, 6
; SI: S_ASHR_I32 {{s[0-9]+}}, [[REG0]], 7
; SI: S_LSHL_B32 [[REG1:s[0-9]+]], {{s[0-9]}}, 6
; SI: S_ASHR_I32 {{s[0-9]+}}, [[REG1]], 7
; EG-NOT: BFE
define void @sext_in_reg_v2i1_in_v2i32_other_amount(<2 x i32> addrspace(1)* %out, <2 x i32> %a, <2 x i32> %b) nounwind {
%c = add <2 x i32> %a, %b
%x = shl <2 x i32> %c, <i32 6, i32 6>
%y = ashr <2 x i32> %x, <i32 7, i32 7>
store <2 x i32> %y, <2 x i32> addrspace(1)* %out, align 2
ret void
}
; FUNC-LABEL: @sext_in_reg_v2i1_to_v2i32
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: BUFFER_STORE_DWORDX2
; EG: BFE
; EG: BFE
define void @sext_in_reg_v2i1_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i32> %a, <2 x i32> %b) nounwind {
%c = add <2 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <2 x i32> %c, <i32 31, i32 31>
%ashr = ashr <2 x i32> %shl, <i32 31, i32 31>
store <2 x i32> %ashr, <2 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_v4i1_to_v4i32
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 1
; SI: BUFFER_STORE_DWORDX4
; EG: BFE
; EG: BFE
; EG: BFE
; EG: BFE
define void @sext_in_reg_v4i1_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i32> %a, <4 x i32> %b) nounwind {
%c = add <4 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <4 x i32> %c, <i32 31, i32 31, i32 31, i32 31>
%ashr = ashr <4 x i32> %shl, <i32 31, i32 31, i32 31, i32 31>
store <4 x i32> %ashr, <4 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_v2i8_to_v2i32
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: BUFFER_STORE_DWORDX2
; EG: BFE
; EG: BFE
define void @sext_in_reg_v2i8_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i32> %a, <2 x i32> %b) nounwind {
%c = add <2 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <2 x i32> %c, <i32 24, i32 24>
%ashr = ashr <2 x i32> %shl, <i32 24, i32 24>
store <2 x i32> %ashr, <2 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_v4i8_to_v4i32
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: BUFFER_STORE_DWORDX4
; EG: BFE
; EG: BFE
; EG: BFE
; EG: BFE
define void @sext_in_reg_v4i8_to_v4i32(<4 x i32> addrspace(1)* %out, <4 x i32> %a, <4 x i32> %b) nounwind {
%c = add <4 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <4 x i32> %c, <i32 24, i32 24, i32 24, i32 24>
%ashr = ashr <4 x i32> %shl, <i32 24, i32 24, i32 24, i32 24>
store <4 x i32> %ashr, <4 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @sext_in_reg_v2i16_to_v2i32
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: V_BFE_I32 {{v[0-9]+}}, {{s[0-9]+}}, 0, 8
; SI: BUFFER_STORE_DWORDX2
; EG: BFE
; EG: BFE
define void @sext_in_reg_v2i16_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i32> %a, <2 x i32> %b) nounwind {
%c = add <2 x i32> %a, %b ; add to prevent folding into extload
%shl = shl <2 x i32> %c, <i32 24, i32 24>
%ashr = ashr <2 x i32> %shl, <i32 24, i32 24>
store <2 x i32> %ashr, <2 x i32> addrspace(1)* %out, align 8
ret void
}
; FUNC-LABEL: @testcase
define void @testcase(i8 addrspace(1)* %out, i8 %a) nounwind {
%and_a_1 = and i8 %a, 1
%cmp_eq = icmp eq i8 %and_a_1, 0
%cmp_slt = icmp slt i8 %a, 0
%sel0 = select i1 %cmp_slt, i8 0, i8 %a
%sel1 = select i1 %cmp_eq, i8 0, i8 %a
%xor = xor i8 %sel0, %sel1
store i8 %xor, i8 addrspace(1)* %out
ret void
}
; FUNC-LABEL: @testcase_3
define void @testcase_3(i8 addrspace(1)* %out, i8 %a) nounwind {
%and_a_1 = and i8 %a, 1
%cmp_eq = icmp eq i8 %and_a_1, 0
%cmp_slt = icmp slt i8 %a, 0
%sel0 = select i1 %cmp_slt, i8 0, i8 %a
%sel1 = select i1 %cmp_eq, i8 0, i8 %a
%xor = xor i8 %sel0, %sel1
store i8 %xor, i8 addrspace(1)* %out
ret void
}
; FIXME: The BFE should really be eliminated. I think it should happen
; when computeMaskedBitsForTargetNode is implemented for imax.
; FUNC-LABEL: @sext_in_reg_to_illegal_type
; SI: BUFFER_LOAD_SBYTE
; SI: V_MAX_I32
; SI: V_BFE_I32
; SI: BUFFER_STORE_SHORT
define void @sext_in_reg_to_illegal_type(i16 addrspace(1)* nocapture %out, i8 addrspace(1)* nocapture %src) nounwind {
%tmp5 = load i8 addrspace(1)* %src, align 1
%tmp2 = sext i8 %tmp5 to i32
%tmp3 = tail call i32 @llvm.AMDGPU.imax(i32 %tmp2, i32 0) nounwind readnone
%tmp4 = trunc i32 %tmp3 to i8
%tmp6 = sext i8 %tmp4 to i16
store i16 %tmp6, i16 addrspace(1)* %out, align 2
ret void
}
|
