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
|
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=tonga -mattr=-flat-for-global < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=gfx900 -mattr=-flat-for-global < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
; The code generated by sdiv is long and complex and may frequently change.
; The goal of this test is to make sure the ISel doesn't fail.
;
; This program was previously failing to compile when one of the selectcc
; opcodes generated by the sdiv lowering was being legalized and optimized to:
; selectcc Remainder -1, 0, -1, SETGT
; This was fixed by adding an additional pattern in R600Instructions.td to
; match this pattern with a CNDGE_INT.
; FUNC-LABEL: {{^}}sdiv_i32:
; EG: CF_END
define amdgpu_kernel void @sdiv_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32, i32 addrspace(1)* %in, i32 1
%num = load i32, i32 addrspace(1) * %in
%den = load i32, i32 addrspace(1) * %den_ptr
%result = sdiv i32 %num, %den
store i32 %result, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}sdiv_i32_4:
define amdgpu_kernel void @sdiv_i32_4(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%num = load i32, i32 addrspace(1) * %in
%result = sdiv i32 %num, 4
store i32 %result, i32 addrspace(1)* %out
ret void
}
; Multiply by a weird constant to make sure setIntDivIsCheap is
; working.
; FUNC-LABEL: {{^}}slow_sdiv_i32_3435:
; SI-DAG: buffer_load_dword [[VAL:v[0-9]+]],
; SI-DAG: v_mov_b32_e32 [[MAGIC:v[0-9]+]], 0x98a1930b
; SI: v_mul_hi_i32 [[TMP:v[0-9]+]], [[VAL]], [[MAGIC]]
; SI: v_add_{{[iu]}}32
; SI: v_lshrrev_b32
; SI: v_ashrrev_i32
; SI: v_add_{{[iu]}}32
; SI: buffer_store_dword
; SI: s_endpgm
define amdgpu_kernel void @slow_sdiv_i32_3435(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%num = load i32, i32 addrspace(1) * %in
%result = sdiv i32 %num, 3435
store i32 %result, i32 addrspace(1)* %out
ret void
}
define amdgpu_kernel void @sdiv_v2i32(<2 x i32> addrspace(1)* %out, <2 x i32> addrspace(1)* %in) {
%den_ptr = getelementptr <2 x i32>, <2 x i32> addrspace(1)* %in, i32 1
%num = load <2 x i32>, <2 x i32> addrspace(1) * %in
%den = load <2 x i32>, <2 x i32> addrspace(1) * %den_ptr
%result = sdiv <2 x i32> %num, %den
store <2 x i32> %result, <2 x i32> addrspace(1)* %out
ret void
}
define amdgpu_kernel void @sdiv_v2i32_4(<2 x i32> addrspace(1)* %out, <2 x i32> addrspace(1)* %in) {
%num = load <2 x i32>, <2 x i32> addrspace(1) * %in
%result = sdiv <2 x i32> %num, <i32 4, i32 4>
store <2 x i32> %result, <2 x i32> addrspace(1)* %out
ret void
}
define amdgpu_kernel void @sdiv_v4i32(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
%den_ptr = getelementptr <4 x i32>, <4 x i32> addrspace(1)* %in, i32 1
%num = load <4 x i32>, <4 x i32> addrspace(1) * %in
%den = load <4 x i32>, <4 x i32> addrspace(1) * %den_ptr
%result = sdiv <4 x i32> %num, %den
store <4 x i32> %result, <4 x i32> addrspace(1)* %out
ret void
}
define amdgpu_kernel void @sdiv_v4i32_4(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
%num = load <4 x i32>, <4 x i32> addrspace(1) * %in
%result = sdiv <4 x i32> %num, <i32 4, i32 4, i32 4, i32 4>
store <4 x i32> %result, <4 x i32> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}v_sdiv_i8:
; SI: v_rcp_f32
; SI: v_bfe_i32 [[BFE:v[0-9]+]], v{{[0-9]+}}, 0, 8
; SI: buffer_store_dword [[BFE]]
define amdgpu_kernel void @v_sdiv_i8(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
%den_ptr = getelementptr i8, i8 addrspace(1)* %in, i8 1
%num = load i8, i8 addrspace(1) * %in
%den = load i8, i8 addrspace(1) * %den_ptr
%result = sdiv i8 %num, %den
%result.ext = sext i8 %result to i32
store i32 %result.ext, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}v_sdiv_i23:
; SI: v_rcp_f32
; SI: v_bfe_i32 [[BFE:v[0-9]+]], v{{[0-9]+}}, 0, 23
; SI: buffer_store_dword [[BFE]]
define amdgpu_kernel void @v_sdiv_i23(i32 addrspace(1)* %out, i23 addrspace(1)* %in) {
%den_ptr = getelementptr i23, i23 addrspace(1)* %in, i23 1
%num = load i23, i23 addrspace(1) * %in
%den = load i23, i23 addrspace(1) * %den_ptr
%result = sdiv i23 %num, %den
%result.ext = sext i23 %result to i32
store i32 %result.ext, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}v_sdiv_i24:
; SI: v_rcp_f32
; SI: v_bfe_i32 [[BFE:v[0-9]+]], v{{[0-9]+}}, 0, 24
; SI: buffer_store_dword [[BFE]]
define amdgpu_kernel void @v_sdiv_i24(i32 addrspace(1)* %out, i24 addrspace(1)* %in) {
%den_ptr = getelementptr i24, i24 addrspace(1)* %in, i24 1
%num = load i24, i24 addrspace(1) * %in
%den = load i24, i24 addrspace(1) * %den_ptr
%result = sdiv i24 %num, %den
%result.ext = sext i24 %result to i32
store i32 %result.ext, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}v_sdiv_i25:
; SI-NOT: v_rcp_f32
define amdgpu_kernel void @v_sdiv_i25(i32 addrspace(1)* %out, i25 addrspace(1)* %in) {
%den_ptr = getelementptr i25, i25 addrspace(1)* %in, i25 1
%num = load i25, i25 addrspace(1) * %in
%den = load i25, i25 addrspace(1) * %den_ptr
%result = sdiv i25 %num, %den
%result.ext = sext i25 %result to i32
store i32 %result.ext, i32 addrspace(1)* %out
ret void
}
; Tests for 64-bit divide bypass.
; define amdgpu_kernel void @test_get_quotient(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
; %result = sdiv i64 %a, %b
; store i64 %result, i64 addrspace(1)* %out, align 8
; ret void
; }
; define amdgpu_kernel void @test_get_remainder(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
; %result = srem i64 %a, %b
; store i64 %result, i64 addrspace(1)* %out, align 8
; ret void
; }
; define amdgpu_kernel void @test_get_quotient_and_remainder(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
; %resultdiv = sdiv i64 %a, %b
; %resultrem = srem i64 %a, %b
; %result = add i64 %resultdiv, %resultrem
; store i64 %result, i64 addrspace(1)* %out, align 8
; ret void
; }
; FUNC-LABEL: @scalarize_mulhs_4xi32
; SI: v_mul_hi_i32
; SI: v_mul_hi_i32
; SI: v_mul_hi_i32
; SI: v_mul_hi_i32
define amdgpu_kernel void @scalarize_mulhs_4xi32(<4 x i32> addrspace(1)* nocapture readonly %in, <4 x i32> addrspace(1)* nocapture %out) {
%1 = load <4 x i32>, <4 x i32> addrspace(1)* %in, align 16
%2 = sdiv <4 x i32> %1, <i32 53668, i32 53668, i32 53668, i32 53668>
store <4 x i32> %2, <4 x i32> addrspace(1)* %out, align 16
ret void
}
|
