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
|
let Namespace = "AMDGPU" in {
def low : SubRegIndex;
def high : SubRegIndex;
def sub0 : SubRegIndex;
def sub1 : SubRegIndex;
def sub2 : SubRegIndex;
def sub3 : SubRegIndex;
def sub4 : SubRegIndex;
def sub5 : SubRegIndex;
def sub6 : SubRegIndex;
def sub7 : SubRegIndex;
}
class SIReg <string n, bits<16> encoding = 0> : Register<n> {
let Namespace = "AMDGPU";
let HWEncoding = encoding;
}
class SI_64 <string n, list<Register> subregs, bits<16> encoding> : RegisterWithSubRegs<n, subregs> {
let Namespace = "AMDGPU";
let SubRegIndices = [low, high];
let HWEncoding = encoding;
}
class SGPR_32 <bits<16> num, string name> : SIReg<name, num>;
class VGPR_32 <bits<16> num, string name> : SIReg<name, num>;
// Special Registers
def VCC : SIReg<"VCC", 106>;
def EXEC_LO : SIReg <"EXEC LO", 126>;
def EXEC_HI : SIReg <"EXEC HI", 127>;
def EXEC : SI_64<"EXEC", [EXEC_LO, EXEC_HI], 126>;
def SCC : SIReg<"SCC", 253>;
def SREG_LIT_0 : SIReg <"S LIT 0", 128>;
def SI_LITERAL_CONSTANT : SIReg<"LITERAL CONSTANT", 255>;
def M0 : SIReg <"M0", 124>;
//Interpolation registers
def PERSP_SAMPLE_I : SIReg <"PERSP_SAMPLE_I">;
def PERSP_SAMPLE_J : SIReg <"PERSP_SAMPLE_J">;
def PERSP_CENTER_I : SIReg <"PERSP_CENTER_I">;
def PERSP_CENTER_J : SIReg <"PERSP_CENTER_J">;
def PERSP_CENTROID_I : SIReg <"PERSP_CENTROID_I">;
def PERSP_CENTROID_J : SIReg <"PERP_CENTROID_J">;
def PERSP_I_W : SIReg <"PERSP_I_W">;
def PERSP_J_W : SIReg <"PERSP_J_W">;
def PERSP_1_W : SIReg <"PERSP_1_W">;
def LINEAR_SAMPLE_I : SIReg <"LINEAR_SAMPLE_I">;
def LINEAR_SAMPLE_J : SIReg <"LINEAR_SAMPLE_J">;
def LINEAR_CENTER_I : SIReg <"LINEAR_CENTER_I">;
def LINEAR_CENTER_J : SIReg <"LINEAR_CENTER_J">;
def LINEAR_CENTROID_I : SIReg <"LINEAR_CENTROID_I">;
def LINEAR_CENTROID_J : SIReg <"LINEAR_CENTROID_J">;
def LINE_STIPPLE_TEX_COORD : SIReg <"LINE_STIPPLE_TEX_COORD">;
def POS_X_FLOAT : SIReg <"POS_X_FLOAT">;
def POS_Y_FLOAT : SIReg <"POS_Y_FLOAT">;
def POS_Z_FLOAT : SIReg <"POS_Z_FLOAT">;
def POS_W_FLOAT : SIReg <"POS_W_FLOAT">;
def FRONT_FACE : SIReg <"FRONT_FACE">;
def ANCILLARY : SIReg <"ANCILLARY">;
def SAMPLE_COVERAGE : SIReg <"SAMPLE_COVERAGE">;
def POS_FIXED_PT : SIReg <"POS_FIXED_PT">;
// SGPR 32-bit registers
foreach Index = 0-101 in {
def SGPR#Index : SGPR_32 <Index, "SGPR"#Index>;
}
def SGPR_32 : RegisterClass<"AMDGPU", [f32, i32], 32,
(add (sequence "SGPR%u", 0, 101))>;
// SGPR 64-bit registers
def SGPR_64 : RegisterTuples<[low, high],
[(add (decimate SGPR_32, 2)),
(add(decimate (rotl SGPR_32, 1), 2))]>;
// SGPR 128-bit registers
def SGPR_128 : RegisterTuples<[sel_x, sel_y, sel_z, sel_w],
[(add (decimate SGPR_32, 4)),
(add (decimate (rotl SGPR_32, 1), 4)),
(add (decimate (rotl SGPR_32, 2), 4)),
(add (decimate (rotl SGPR_32, 3), 4))]>;
// SGPR 256-bit registers
def SGPR_256 : RegisterTuples<[sub0, sub1, sub2, sub3, sub4, sub5, sub6, sub7],
[(add (decimate SGPR_32, 8)),
(add (decimate (rotl SGPR_32, 1), 8)),
(add (decimate (rotl SGPR_32, 2), 8)),
(add (decimate (rotl SGPR_32, 3), 8)),
(add (decimate (rotl SGPR_32, 4), 8)),
(add (decimate (rotl SGPR_32, 5), 8)),
(add (decimate (rotl SGPR_32, 6), 8)),
(add (decimate (rotl SGPR_32, 7), 8))]>;
// VGPR 32-bit registers
foreach Index = 0-255 in {
def VGPR#Index : VGPR_32 <Index, "VGPR"#Index>;
}
def VGPR_32 : RegisterClass<"AMDGPU", [f32, i32], 32,
(add (sequence "VGPR%u", 0, 255))>;
// VGPR 64-bit registers
def VGPR_64 : RegisterTuples<[low, high],
[(add VGPR_32),
(add (rotl VGPR_32, 1))]>;
// VGPR 128-bit registers
def VGPR_128 : RegisterTuples<[sel_x, sel_y, sel_z, sel_w],
[(add VGPR_32),
(add (rotl VGPR_32, 1)),
(add (rotl VGPR_32, 2)),
(add (rotl VGPR_32, 3))]>;
// Register class for all scalar registers (SGPRs + Special Registers)
def SReg_32 : RegisterClass<"AMDGPU", [f32, i32], 32,
(add SGPR_32, SREG_LIT_0, M0, EXEC_LO, EXEC_HI)
>;
def SReg_64 : RegisterClass<"AMDGPU", [i64], 64, (add SGPR_64, VCC, EXEC)>;
def SReg_1 : RegisterClass<"AMDGPU", [i1], 1, (add VCC, SGPR_64, EXEC)>;
def SReg_128 : RegisterClass<"AMDGPU", [v4f32, v4i32], 128, (add SGPR_128)>;
def SReg_256 : RegisterClass<"AMDGPU", [v8i32], 256, (add SGPR_256)>;
// Register class for all vector registers (VGPRs + Interploation Registers)
def VReg_32 : RegisterClass<"AMDGPU", [f32, i32], 32,
(add VGPR_32,
PERSP_SAMPLE_I, PERSP_SAMPLE_J,
PERSP_CENTER_I, PERSP_CENTER_J,
PERSP_CENTROID_I, PERSP_CENTROID_J,
PERSP_I_W, PERSP_J_W, PERSP_1_W,
LINEAR_SAMPLE_I, LINEAR_SAMPLE_J,
LINEAR_CENTER_I, LINEAR_CENTER_J,
LINEAR_CENTROID_I, LINEAR_CENTROID_J,
LINE_STIPPLE_TEX_COORD,
POS_X_FLOAT,
POS_Y_FLOAT,
POS_Z_FLOAT,
POS_W_FLOAT,
FRONT_FACE,
ANCILLARY,
SAMPLE_COVERAGE,
POS_FIXED_PT
)
>;
def VReg_64 : RegisterClass<"AMDGPU", [i64], 64, (add VGPR_64)>;
def VReg_128 : RegisterClass<"AMDGPU", [v4f32], 128, (add VGPR_128)>;
// AllReg_* - A set of all scalar and vector registers of a given width.
def AllReg_32 : RegisterClass<"AMDGPU", [f32, i32], 32, (add VReg_32, SReg_32)>;
def AllReg_64 : RegisterClass<"AMDGPU", [f64, i64], 64, (add SReg_64, VReg_64)>;
// Special register classes for predicates and the M0 register
def SCCReg : RegisterClass<"AMDGPU", [i1], 1, (add SCC)>;
def VCCReg : RegisterClass<"AMDGPU", [i1], 1, (add VCC)>;
def EXECReg : RegisterClass<"AMDGPU", [i1], 1, (add EXEC)>;
def M0Reg : RegisterClass<"AMDGPU", [i32], 32, (add M0)>;
|
