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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx704 < %s | FileCheck -check-prefix=GFX7 %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefix=GFX9 %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 < %s | FileCheck -check-prefix=GFX11 %s
define i32 @s_add_co_select_user() {
; GFX7-LABEL: s_add_co_select_user:
; GFX7: ; %bb.0: ; %bb
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: s_mov_b64 s[4:5], 0
; GFX7-NEXT: s_load_dword s6, s[4:5], 0x0
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e64 v0, s[4:5], s6, s6
; GFX7-NEXT: s_or_b32 s4, s4, s5
; GFX7-NEXT: s_cmp_lg_u32 s4, 0
; GFX7-NEXT: s_addc_u32 s7, s6, 0
; GFX7-NEXT: s_cselect_b64 s[4:5], -1, 0
; GFX7-NEXT: s_and_b64 s[4:5], s[4:5], exec
; GFX7-NEXT: s_cselect_b32 s4, s7, 0
; GFX7-NEXT: s_cmp_gt_u32 s6, 31
; GFX7-NEXT: v_mov_b32_e32 v1, s4
; GFX7-NEXT: s_cselect_b64 vcc, -1, 0
; GFX7-NEXT: v_cndmask_b32_e32 v0, v1, v0, vcc
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: s_add_co_select_user:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: s_mov_b64 s[4:5], 0
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_add_co_u32_e64 v0, s[4:5], s6, s6
; GFX9-NEXT: s_cmp_lg_u64 s[4:5], 0
; GFX9-NEXT: s_addc_u32 s7, s6, 0
; GFX9-NEXT: s_cselect_b64 s[4:5], -1, 0
; GFX9-NEXT: s_and_b64 s[4:5], s[4:5], exec
; GFX9-NEXT: s_cselect_b32 s4, s7, 0
; GFX9-NEXT: s_cmp_gt_u32 s6, 31
; GFX9-NEXT: v_mov_b32_e32 v1, s4
; GFX9-NEXT: s_cselect_b64 vcc, -1, 0
; GFX9-NEXT: v_cndmask_b32_e32 v0, v1, v0, vcc
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: s_add_co_select_user:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: s_mov_b64 s[4:5], 0
; GFX10-NEXT: s_load_dword s4, s[4:5], 0x0
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: v_add_co_u32 v0, s5, s4, s4
; GFX10-NEXT: s_cmp_lg_u32 s5, 0
; GFX10-NEXT: s_addc_u32 s5, s4, 0
; GFX10-NEXT: s_cselect_b32 s6, -1, 0
; GFX10-NEXT: s_and_b32 s6, s6, exec_lo
; GFX10-NEXT: s_cselect_b32 s5, s5, 0
; GFX10-NEXT: s_cmp_gt_u32 s4, 31
; GFX10-NEXT: s_cselect_b32 vcc_lo, -1, 0
; GFX10-NEXT: v_cndmask_b32_e32 v0, s5, v0, vcc_lo
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: s_add_co_select_user:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: s_mov_b64 s[0:1], 0
; GFX11-NEXT: s_load_b32 s0, s[0:1], 0x0
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
; GFX11-NEXT: v_add_co_u32 v0, s1, s0, s0
; GFX11-NEXT: s_cmp_lg_u32 s1, 0
; GFX11-NEXT: s_addc_u32 s1, s0, 0
; GFX11-NEXT: s_cselect_b32 s2, -1, 0
; GFX11-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX11-NEXT: s_and_b32 s2, s2, exec_lo
; GFX11-NEXT: s_cselect_b32 s1, s1, 0
; GFX11-NEXT: s_cmp_gt_u32 s0, 31
; GFX11-NEXT: s_cselect_b32 vcc_lo, -1, 0
; GFX11-NEXT: v_cndmask_b32_e32 v0, s1, v0, vcc_lo
; GFX11-NEXT: s_setpc_b64 s[30:31]
bb:
%i = load volatile i32, ptr addrspace(4) null, align 8
%i1 = add i32 %i, %i
%i2 = icmp ult i32 %i1, %i
%i3 = zext i1 %i2 to i32
%i4 = add nuw nsw i32 %i3, 0
%i5 = add i32 %i4, %i
%i6 = icmp ult i32 %i5, %i4
%i7 = select i1 %i6, i32 %i5, i32 0
%i8 = icmp ugt i32 %i, 31
%i9 = select i1 %i8, i32 %i1, i32 %i7
ret i32 %i9
}
define amdgpu_kernel void @s_add_co_br_user(i32 %i) {
; GFX7-LABEL: s_add_co_br_user:
; GFX7: ; %bb.0: ; %bb
; GFX7-NEXT: s_load_dword s2, s[8:9], 0x0
; GFX7-NEXT: s_add_i32 s12, s12, s17
; GFX7-NEXT: s_lshr_b32 flat_scratch_hi, s12, 8
; GFX7-NEXT: s_mov_b32 flat_scratch_lo, s13
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: s_add_i32 s0, s2, s2
; GFX7-NEXT: s_cmp_lt_u32 s0, s2
; GFX7-NEXT: s_cselect_b64 s[0:1], -1, 0
; GFX7-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[0:1]
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: s_cmp_lg_u32 s0, 0
; GFX7-NEXT: s_addc_u32 s0, s2, 0
; GFX7-NEXT: v_cmp_ge_u32_e32 vcc, s0, v0
; GFX7-NEXT: s_cbranch_vccnz .LBB1_2
; GFX7-NEXT: ; %bb.1: ; %bb0
; GFX7-NEXT: v_mov_b32_e32 v0, 0
; GFX7-NEXT: v_mov_b32_e32 v1, 0
; GFX7-NEXT: v_mov_b32_e32 v2, 9
; GFX7-NEXT: flat_store_dword v[0:1], v2
; GFX7-NEXT: s_waitcnt vmcnt(0)
; GFX7-NEXT: .LBB1_2: ; %bb1
; GFX7-NEXT: v_mov_b32_e32 v0, 0
; GFX7-NEXT: v_mov_b32_e32 v1, 0
; GFX7-NEXT: v_mov_b32_e32 v2, 10
; GFX7-NEXT: flat_store_dword v[0:1], v2
; GFX7-NEXT: s_waitcnt vmcnt(0)
; GFX7-NEXT: s_endpgm
;
; GFX9-LABEL: s_add_co_br_user:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_load_dword s2, s[8:9], 0x0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_add_i32 s0, s2, s2
; GFX9-NEXT: s_cmp_lt_u32 s0, s2
; GFX9-NEXT: s_cselect_b64 s[0:1], -1, 0
; GFX9-NEXT: s_cmp_lg_u64 s[0:1], 0
; GFX9-NEXT: v_cndmask_b32_e64 v0, 0, 1, s[0:1]
; GFX9-NEXT: s_addc_u32 s0, s2, 0
; GFX9-NEXT: v_cmp_ge_u32_e32 vcc, s0, v0
; GFX9-NEXT: s_cbranch_vccnz .LBB1_2
; GFX9-NEXT: ; %bb.1: ; %bb0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: v_mov_b32_e32 v1, 0
; GFX9-NEXT: v_mov_b32_e32 v2, 9
; GFX9-NEXT: global_store_dword v[0:1], v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: .LBB1_2: ; %bb1
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: v_mov_b32_e32 v1, 0
; GFX9-NEXT: v_mov_b32_e32 v2, 10
; GFX9-NEXT: global_store_dword v[0:1], v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: s_add_co_br_user:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_load_dword s0, s[8:9], 0x0
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: s_add_i32 s1, s0, s0
; GFX10-NEXT: s_cmp_lt_u32 s1, s0
; GFX10-NEXT: s_cselect_b32 s1, -1, 0
; GFX10-NEXT: v_cndmask_b32_e64 v0, 0, 1, s1
; GFX10-NEXT: s_cmp_lg_u32 s1, 0
; GFX10-NEXT: s_addc_u32 s0, s0, 0
; GFX10-NEXT: v_cmp_ge_u32_e32 vcc_lo, s0, v0
; GFX10-NEXT: s_cbranch_vccnz .LBB1_2
; GFX10-NEXT: ; %bb.1: ; %bb0
; GFX10-NEXT: v_mov_b32_e32 v0, 0
; GFX10-NEXT: v_mov_b32_e32 v1, 0
; GFX10-NEXT: v_mov_b32_e32 v2, 9
; GFX10-NEXT: global_store_dword v[0:1], v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: .LBB1_2: ; %bb1
; GFX10-NEXT: v_mov_b32_e32 v0, 0
; GFX10-NEXT: v_mov_b32_e32 v1, 0
; GFX10-NEXT: v_mov_b32_e32 v2, 10
; GFX10-NEXT: global_store_dword v[0:1], v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: s_endpgm
;
; GFX11-LABEL: s_add_co_br_user:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_load_b32 s0, s[4:5], 0x0
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
; GFX11-NEXT: s_add_i32 s1, s0, s0
; GFX11-NEXT: s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
; GFX11-NEXT: s_cmp_lt_u32 s1, s0
; GFX11-NEXT: s_cselect_b32 s1, -1, 0
; GFX11-NEXT: v_cndmask_b32_e64 v0, 0, 1, s1
; GFX11-NEXT: s_cmp_lg_u32 s1, 0
; GFX11-NEXT: s_addc_u32 s0, s0, 0
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_1) | instid1(SALU_CYCLE_1)
; GFX11-NEXT: v_cmp_ge_u32_e32 vcc_lo, s0, v0
; GFX11-NEXT: s_cbranch_vccnz .LBB1_2
; GFX11-NEXT: ; %bb.1: ; %bb0
; GFX11-NEXT: v_mov_b32_e32 v0, 0
; GFX11-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 9
; GFX11-NEXT: global_store_b32 v[0:1], v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: .LBB1_2: ; %bb1
; GFX11-NEXT: v_mov_b32_e32 v0, 0
; GFX11-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 10
; GFX11-NEXT: global_store_b32 v[0:1], v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: s_endpgm
bb:
%i1 = add i32 %i, %i
%i2 = icmp ult i32 %i1, %i
%i3 = zext i1 %i2 to i32
%i4 = add nuw nsw i32 %i3, 0
%i5 = add i32 %i4, %i
%i6 = icmp ult i32 %i5, %i4
%i7 = select i1 %i6, i32 %i5, i32 0
br i1 %i6, label %bb0, label %bb1
bb0:
store volatile i32 9, ptr addrspace(1) null
br label %bb1
bb1:
store volatile i32 10, ptr addrspace(1) null
ret void
}
|
