RISCVSchedBullet.td
7.71 KB
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
//==- RISCVSchedBullet.td - Bullet Scheduling Definitions ----*- tablegen -*-=//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// The following definitions describe the simpler per-operand machine model.
// This works with MachineScheduler. See MCSchedule.h for details.
// Bullet machine model for scheduling and other instruction cost heuristics.
def BulletModel : SchedMachineModel {
let MicroOpBufferSize = 0; // Explicitly set to zero since Bullet is in-order.
let IssueWidth = 2; // 2 micro-ops are dispatched per cycle.
let LoadLatency = 3;
let MispredictPenalty = 3;
let CompleteModel = 0;
let UnsupportedFeatures = [HasStdExtV, HasStdExtZvamo, HasStdExtZvlsseg];
}
// The Bullet microarchitecure has two pipelines: A and B.
// Pipe A can handle memory, integer alu and vector operations.
// Pipe B can handle integer alu, control flow, integer multiply and divide,
// and floating point computation.
let SchedModel = BulletModel in {
let BufferSize = 0 in {
def BulletPipeA : ProcResource<1>;
def BulletPipeB : ProcResource<1>;
}
let BufferSize = 1 in {
def BulletIDiv : ProcResource<1> { let Super = BulletPipeB; } // Int Division
def BulletFDiv : ProcResource<1> { let Super = BulletPipeB; } // FP Division/Sqrt
}
def BulletPipeAB : ProcResGroup<[BulletPipeA, BulletPipeB]>;
// Branching
def : WriteRes<WriteJmp, [BulletPipeB]>;
def : WriteRes<WriteJal, [BulletPipeB]>;
def : WriteRes<WriteJalr, [BulletPipeB]>;
def : WriteRes<WriteJmpReg, [BulletPipeB]>;
// Integer arithmetic and logic
let Latency = 3 in {
def : WriteRes<WriteIALU, [BulletPipeAB]>;
def : WriteRes<WriteIALU32, [BulletPipeAB]>;
def : WriteRes<WriteShift, [BulletPipeAB]>;
def : WriteRes<WriteShift32, [BulletPipeAB]>;
}
// Integer multiplication
let Latency = 3 in {
def : WriteRes<WriteIMul, [BulletPipeB]>;
def : WriteRes<WriteIMul32, [BulletPipeB]>;
}
// Integer division
def : WriteRes<WriteIDiv, [BulletPipeB, BulletIDiv]> {
let Latency = 16;
let ResourceCycles = [1, 15];
}
def : WriteRes<WriteIDiv32, [BulletPipeB, BulletIDiv]> {
let Latency = 16;
let ResourceCycles = [1, 15];
}
// Memory
def : WriteRes<WriteSTB, [BulletPipeA]>;
def : WriteRes<WriteSTH, [BulletPipeA]>;
def : WriteRes<WriteSTW, [BulletPipeA]>;
def : WriteRes<WriteSTD, [BulletPipeA]>;
def : WriteRes<WriteFST32, [BulletPipeA]>;
def : WriteRes<WriteFST64, [BulletPipeA]>;
let Latency = 3 in {
def : WriteRes<WriteLDB, [BulletPipeA]>;
def : WriteRes<WriteLDH, [BulletPipeA]>;
def : WriteRes<WriteLDW, [BulletPipeA]>;
def : WriteRes<WriteLDWU, [BulletPipeA]>;
def : WriteRes<WriteLDD, [BulletPipeA]>;
}
let Latency = 2 in {
def : WriteRes<WriteFLD32, [BulletPipeA]>;
def : WriteRes<WriteFLD64, [BulletPipeA]>;
}
// Atomic memory
def : WriteRes<WriteAtomicSTW, [BulletPipeA]>;
def : WriteRes<WriteAtomicSTD, [BulletPipeA]>;
let Latency = 3 in {
def : WriteRes<WriteAtomicW, [BulletPipeA]>;
def : WriteRes<WriteAtomicD, [BulletPipeA]>;
def : WriteRes<WriteAtomicLDW, [BulletPipeA]>;
def : WriteRes<WriteAtomicLDD, [BulletPipeA]>;
}
// Single precision.
let Latency = 5 in {
def : WriteRes<WriteFALU32, [BulletPipeB]>;
def : WriteRes<WriteFMul32, [BulletPipeB]>;
def : WriteRes<WriteFMulAdd32, [BulletPipeB]>;
def : WriteRes<WriteFMulSub32, [BulletPipeB]>;
}
let Latency = 3 in {
def : WriteRes<WriteFSGNJ32, [BulletPipeB]>;
def : WriteRes<WriteFMinMax32, [BulletPipeB]>;
}
def : WriteRes<WriteFDiv32, [BulletPipeB, BulletFDiv]> { let Latency = 27;
let ResourceCycles = [1, 26]; }
def : WriteRes<WriteFSqrt32, [BulletPipeB, BulletFDiv]> { let Latency = 27;
let ResourceCycles = [1, 26]; }
// Double precision
let Latency = 7 in {
def : WriteRes<WriteFALU64, [BulletPipeB]>;
def : WriteRes<WriteFMul64, [BulletPipeB]>;
def : WriteRes<WriteFMulAdd64, [BulletPipeB]>;
def : WriteRes<WriteFMulSub64, [BulletPipeB]>;
}
let Latency = 3 in {
def : WriteRes<WriteFSGNJ64, [BulletPipeB]>;
def : WriteRes<WriteFMinMax64, [BulletPipeB]>;
}
def : WriteRes<WriteFDiv64, [BulletPipeB, BulletFDiv]> { let Latency = 56;
let ResourceCycles = [1, 55]; }
def : WriteRes<WriteFSqrt64, [BulletPipeB, BulletFDiv]> { let Latency = 56;
let ResourceCycles = [1, 55]; }
// Conversions
let Latency = 3 in {
def : WriteRes<WriteFCvtI32ToF32, [BulletPipeB]>;
def : WriteRes<WriteFCvtI32ToF64, [BulletPipeB]>;
def : WriteRes<WriteFCvtI64ToF32, [BulletPipeB]>;
def : WriteRes<WriteFCvtI64ToF64, [BulletPipeB]>;
def : WriteRes<WriteFCvtF32ToI32, [BulletPipeB]>;
def : WriteRes<WriteFCvtF32ToI64, [BulletPipeB]>;
def : WriteRes<WriteFCvtF32ToF64, [BulletPipeB]>;
def : WriteRes<WriteFCvtF64ToI32, [BulletPipeB]>;
def : WriteRes<WriteFCvtF64ToI64, [BulletPipeB]>;
def : WriteRes<WriteFCvtF64ToF32, [BulletPipeB]>;
def : WriteRes<WriteFClass32, [BulletPipeB]>;
def : WriteRes<WriteFClass64, [BulletPipeB]>;
def : WriteRes<WriteFCmp32, [BulletPipeB]>;
def : WriteRes<WriteFCmp64, [BulletPipeB]>;
def : WriteRes<WriteFMovI32ToF32, [BulletPipeB]>;
def : WriteRes<WriteFMovF32ToI32, [BulletPipeB]>;
def : WriteRes<WriteFMovI64ToF64, [BulletPipeB]>;
def : WriteRes<WriteFMovF64ToI64, [BulletPipeB]>;
}
// Others
def : WriteRes<WriteCSR, [BulletPipeB]>;
def : WriteRes<WriteNop, []>;
def : InstRW<[WriteIALU], (instrs COPY)>;
//===----------------------------------------------------------------------===//
// Bypass and advance
def : ReadAdvance<ReadJmp, 0>;
def : ReadAdvance<ReadJalr, 0>;
def : ReadAdvance<ReadCSR, 0>;
def : ReadAdvance<ReadStoreData, 0>;
def : ReadAdvance<ReadMemBase, 0>;
def : ReadAdvance<ReadIALU, 0>;
def : ReadAdvance<ReadIALU32, 0>;
def : ReadAdvance<ReadShift, 0>;
def : ReadAdvance<ReadShift32, 0>;
def : ReadAdvance<ReadIDiv, 0>;
def : ReadAdvance<ReadIDiv32, 0>;
def : ReadAdvance<ReadIMul, 0>;
def : ReadAdvance<ReadIMul32, 0>;
def : ReadAdvance<ReadAtomicWA, 0>;
def : ReadAdvance<ReadAtomicWD, 0>;
def : ReadAdvance<ReadAtomicDA, 0>;
def : ReadAdvance<ReadAtomicDD, 0>;
def : ReadAdvance<ReadAtomicLDW, 0>;
def : ReadAdvance<ReadAtomicLDD, 0>;
def : ReadAdvance<ReadAtomicSTW, 0>;
def : ReadAdvance<ReadAtomicSTD, 0>;
def : ReadAdvance<ReadFMemBase, 0>;
def : ReadAdvance<ReadFALU32, 0>;
def : ReadAdvance<ReadFALU64, 0>;
def : ReadAdvance<ReadFMul32, 0>;
def : ReadAdvance<ReadFMulAdd32, 0>;
def : ReadAdvance<ReadFMulSub32, 0>;
def : ReadAdvance<ReadFMul64, 0>;
def : ReadAdvance<ReadFMulAdd64, 0>;
def : ReadAdvance<ReadFMulSub64, 0>;
def : ReadAdvance<ReadFDiv32, 0>;
def : ReadAdvance<ReadFDiv64, 0>;
def : ReadAdvance<ReadFSqrt32, 0>;
def : ReadAdvance<ReadFSqrt64, 0>;
def : ReadAdvance<ReadFCmp32, 0>;
def : ReadAdvance<ReadFCmp64, 0>;
def : ReadAdvance<ReadFSGNJ32, 0>;
def : ReadAdvance<ReadFSGNJ64, 0>;
def : ReadAdvance<ReadFMinMax32, 0>;
def : ReadAdvance<ReadFMinMax64, 0>;
def : ReadAdvance<ReadFCvtF32ToI32, 0>;
def : ReadAdvance<ReadFCvtF32ToI64, 0>;
def : ReadAdvance<ReadFCvtF64ToI32, 0>;
def : ReadAdvance<ReadFCvtF64ToI64, 0>;
def : ReadAdvance<ReadFCvtI32ToF32, 0>;
def : ReadAdvance<ReadFCvtI32ToF64, 0>;
def : ReadAdvance<ReadFCvtI64ToF32, 0>;
def : ReadAdvance<ReadFCvtI64ToF64, 0>;
def : ReadAdvance<ReadFCvtF32ToF64, 0>;
def : ReadAdvance<ReadFCvtF64ToF32, 0>;
def : ReadAdvance<ReadFMovF32ToI32, 0>;
def : ReadAdvance<ReadFMovI32ToF32, 0>;
def : ReadAdvance<ReadFMovF64ToI64, 0>;
def : ReadAdvance<ReadFMovI64ToF64, 0>;
def : ReadAdvance<ReadFClass32, 0>;
def : ReadAdvance<ReadFClass64, 0>;
}