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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -o - -amdgpu-codegenprepare-mul24=0 < %s | FileCheck -check-prefix=GFX9 %s
define i16 @num_sign_bits_mul_i48_0(i8 %X, i8 %Y, i8 %Z, i8 %W) {
; GFX9-LABEL: num_sign_bits_mul_i48_0:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mul_i32_i24_sdwa v0, sext(v0), sext(v1) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0
; GFX9-NEXT: v_mul_i32_i24_sdwa v1, sext(v2), sext(v3) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0
; GFX9-NEXT: v_mul_i32_i24_e32 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
%A = sext i8 %X to i48
%B = sext i8 %Y to i48
%C = sext i8 %Z to i48
%D = sext i8 %W to i48
%mul0 = mul i48 %A, %B
%mul1 = mul i48 %C, %D
%mul2 = mul i48 %mul0, %mul1
%trunc = trunc i48 %mul2 to i16
ret i16 %trunc
}
define i16 @num_sign_bits_mul_i48_1(i8 %X, i8 %Y, i8 %Z, i8 %W) {
; GFX9-LABEL: num_sign_bits_mul_i48_1:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mul_i32_i24_sdwa v0, sext(v0), sext(v1) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0
; GFX9-NEXT: v_mul_i32_i24_sdwa v2, sext(v2), sext(v3) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0
; GFX9-NEXT: v_mul_hi_i32_i24_e32 v1, v0, v2
; GFX9-NEXT: v_mul_i32_i24_e32 v0, v0, v2
; GFX9-NEXT: v_lshrrev_b64 v[0:1], 24, v[0:1]
; GFX9-NEXT: s_setpc_b64 s[30:31]
%A = sext i8 %X to i48
%B = sext i8 %Y to i48
%C = sext i8 %Z to i48
%D = sext i8 %W to i48
%mul0 = mul i48 %A, %B
%mul1 = mul i48 %C, %D
%mul2 = mul i48 %mul0, %mul1
%ashr = ashr i48 %mul2, 24
%trunc = trunc i48 %ashr to i16
ret i16 %trunc
}
define i32 @num_sign_bits_mul_i32_7(i32 %x, i32 %y, i32 %z, i32 %w) {
; GFX9-LABEL: num_sign_bits_mul_i32_7:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_bfe_i32 v0, v0, 0, 25
; GFX9-NEXT: v_bfe_i32 v1, v1, 0, 25
; GFX9-NEXT: v_bfe_i32 v2, v2, 0, 25
; GFX9-NEXT: v_bfe_i32 v3, v3, 0, 25
; GFX9-NEXT: v_mul_lo_u32 v0, v0, v1
; GFX9-NEXT: v_mul_lo_u32 v1, v2, v3
; GFX9-NEXT: v_mul_lo_u32 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
%x.shl = shl i32 %x, 7
%x.bits = ashr i32 %x.shl, 7
%y.shl = shl i32 %y, 7
%y.bits = ashr i32 %y.shl, 7
%z.shl = shl i32 %z, 7
%z.bits = ashr i32 %z.shl, 7
%w.shl = shl i32 %w, 7
%w.bits = ashr i32 %w.shl, 7
%mul0 = mul i32 %x.bits, %y.bits
%mul1 = mul i32 %z.bits, %w.bits
%mul2 = mul i32 %mul0, %mul1
ret i32 %mul2
}
define i32 @num_sign_bits_mul_i32_8(i32 %x, i32 %y, i32 %z, i32 %w) {
; GFX9-LABEL: num_sign_bits_mul_i32_8:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mul_i32_i24_e32 v0, v0, v1
; GFX9-NEXT: v_mul_i32_i24_e32 v1, v2, v3
; GFX9-NEXT: v_mul_lo_u32 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
%x.shl = shl i32 %x, 8
%x.bits = ashr i32 %x.shl, 8
%y.shl = shl i32 %y, 8
%y.bits = ashr i32 %y.shl, 8
%z.shl = shl i32 %z, 8
%z.bits = ashr i32 %z.shl, 8
%w.shl = shl i32 %w, 8
%w.bits = ashr i32 %w.shl, 8
%mul0 = mul i32 %x.bits, %y.bits
%mul1 = mul i32 %z.bits, %w.bits
%mul2 = mul i32 %mul0, %mul1
ret i32 %mul2
}
define i32 @num_sign_bits_mul_i32_9(i32 %x, i32 %y, i32 %z, i32 %w) {
; GFX9-LABEL: num_sign_bits_mul_i32_9:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_bfe_i32 v0, v0, 0, 23
; GFX9-NEXT: v_bfe_i32 v1, v1, 0, 23
; GFX9-NEXT: v_bfe_i32 v2, v2, 0, 23
; GFX9-NEXT: v_bfe_i32 v3, v3, 0, 23
; GFX9-NEXT: v_mul_i32_i24_e32 v0, v0, v1
; GFX9-NEXT: v_mul_i32_i24_e32 v1, v2, v3
; GFX9-NEXT: v_mul_lo_u32 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
%x.shl = shl i32 %x, 9
%x.bits = ashr i32 %x.shl, 9
%y.shl = shl i32 %y, 9
%y.bits = ashr i32 %y.shl, 9
%z.shl = shl i32 %z, 9
%z.bits = ashr i32 %z.shl, 9
%w.shl = shl i32 %w, 9
%w.bits = ashr i32 %w.shl, 9
%mul0 = mul i32 %x.bits, %y.bits
%mul1 = mul i32 %z.bits, %w.bits
%mul2 = mul i32 %mul0, %mul1
ret i32 %mul2
}
define i32 @num_sign_bits_mul_i32_10(i32 %x, i32 %y, i32 %z, i32 %w) {
; GFX9-LABEL: num_sign_bits_mul_i32_10:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_bfe_i32 v0, v0, 0, 22
; GFX9-NEXT: v_bfe_i32 v1, v1, 0, 22
; GFX9-NEXT: v_bfe_i32 v2, v2, 0, 22
; GFX9-NEXT: v_bfe_i32 v3, v3, 0, 22
; GFX9-NEXT: v_mul_i32_i24_e32 v0, v0, v1
; GFX9-NEXT: v_mul_i32_i24_e32 v1, v2, v3
; GFX9-NEXT: v_mul_lo_u32 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
%x.shl = shl i32 %x, 10
%x.bits = ashr i32 %x.shl, 10
%y.shl = shl i32 %y, 10
%y.bits = ashr i32 %y.shl, 10
%z.shl = shl i32 %z, 10
%z.bits = ashr i32 %z.shl, 10
%w.shl = shl i32 %w, 10
%w.bits = ashr i32 %w.shl, 10
%mul0 = mul i32 %x.bits, %y.bits
%mul1 = mul i32 %z.bits, %w.bits
%mul2 = mul i32 %mul0, %mul1
ret i32 %mul2
}
define i32 @known_bits_mul24() {
; GFX9-LABEL: known_bits_mul24:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_setpc_b64 s[30:31]
%r0 = call i32 @llvm.amdgcn.mul.i24(i32 0, i32 -7)
%r1 = shl i32 %r0, 2
ret i32 %r1
}
declare i32 @llvm.amdgcn.mul.i24(i32, i32)
|
