OpenCores
URL https://opencores.org/ocsvn/openarty/openarty/trunk

Subversion Repositories openarty

[/] [openarty/] [trunk/] [rtl/] [cpu/] [cpudefs.v] - Blame information for rev 30

Go to most recent revision | Details | Compare with Previous | View Log

Line No. Rev Author Line
1 3 dgisselq
///////////////////////////////////////////////////////////////////////////////
2
//
3
// Filename:    cpudefs.v
4
//
5
// Project:     OpenArty, an entirely open SoC based upon the Arty platform
6
//
7
// Purpose:     Some architectures have some needs, others have other needs.
8
//              Some of my projects need a Zip CPU with pipelining, others
9
//      can't handle the timing required to get the answer from the ALU
10
//      back into the input for the ALU.  As each different projects has
11
//      different needs, I can either 1) reconfigure my entire baseline prior
12
//      to building each project, or 2) host a configuration file which contains
13
//      the information regarding each baseline.  This file is that
14
//      configuration file.  It controls how the CPU (not the system,
15
//      peripherals, or other) is defined and implemented.  Several options
16
//      are available within here, making the Zip CPU pipelined or not,
17
//      able to handle a faster clock with more stalls or a slower clock with
18
//      no stalls, etc.
19
//
20
//      This file encapsulates those control options.
21
//
22
//      The number of LUTs the Zip CPU uses varies dramatically with the
23
//      options defined in this file.
24
//
25
//
26
// OpenArty comments:
27
//      My goal on the OpenArty is going to be using the CPU to its fullest
28
//      extent.  All features shall be turned on if they exist, full pipelines,
29
//      multiplies, divides, and hopefully even the 200MHz clock.  This file
30
//      reflects that purpose.
31
//
32
//
33
// Creator:     Dan Gisselquist, Ph.D.
34
//              Gisselquist Technology, LLC
35
//
36
///////////////////////////////////////////////////////////////////////////////
37
//
38
// Copyright (C) 2015-2016, Gisselquist Technology, LLC
39
//
40
// This program is free software (firmware): you can redistribute it and/or
41
// modify it under the terms of  the GNU General Public License as published
42
// by the Free Software Foundation, either version 3 of the License, or (at
43
// your option) any later version.
44
//
45
// This program is distributed in the hope that it will be useful, but WITHOUT
46
// ANY WARRANTY; without even the implied warranty of MERCHANTIBILITY or
47
// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
48
// for more details.
49
//
50
// License:     GPL, v3, as defined and found on www.gnu.org,
51
//              http://www.gnu.org/licenses/gpl.html
52
//
53
//
54
///////////////////////////////////////////////////////////////////////////////
55
`ifndef CPUDEFS_H
56
`define CPUDEFS_H
57
//
58
//
59
// The first couple options control the Zip CPU instruction set, and how
60
// it handles various instructions within the set:
61
//
62
//
63
// OPT_ILLEGAL_INSTRUCTION is part of a new section of code that is supposed
64
// to recognize illegal instructions and interrupt the CPU whenever one such
65
// instruction is encountered.  The goal is to create a soft floating point
66
// unit via this approach, that can then be replaced with a true floating point
67
// unit.  As I'm not there yet, it just catches illegal instructions and
68
// interrupts the CPU on any such instruction--when defined.  Otherwise,
69
// illegal instructions are quietly ignored and their behaviour is ...
70
// undefined. (Many get treated like NOOPs ...)
71
//
72
// I recommend setting this flag, although it can be taken out if area is
73
// critical ...
74
//
75
`define OPT_ILLEGAL_INSTRUCTION
76
//
77
//
78
//
79
// OPT_MULTIPLY controls whether or not the multiply is built and included
80
// in the ALU by default.  Set this option and a parameter will be set that
81
// includes the multiply.  (This parameter may still be overridden, as with
82
// any parameter ...)  If the multiply is not included and
83
// OPT_ILLEGAL_INSTRUCTION is set, then the multiply will create an illegal
84
// instruction that will then trip the illegal instruction trap.
85
//
86
//
87
`define OPT_MULTIPLY    1
88
//
89
//
90
//
91
// OPT_DIVIDE controls whether or not the divide instruction is built and
92
// included into the ZipCPU by default.  Set this option and a parameter will
93
// be set that causes the divide unit to be included.  (This parameter may
94
// still be overridden, as with any parameter ...)  If the divide is not
95
// included and OPT_ILLEGAL_INSTRUCTION is set, then the multiply will create
96
// an illegal instruction exception that will send the CPU into supervisor
97
// mode.
98
//
99
//
100 30 dgisselq
`define OPT_DIVIDE
101 3 dgisselq
//
102
//
103
//
104
// OPT_IMPLEMENT_FPU will (one day) control whether or not the floating point
105
// unit (once I have one) is built and included into the ZipCPU by default. 
106
// At that time, if this option is set then a parameter will be set that
107
// causes the floating point unit to be included.  (This parameter may
108
// still be overridden, as with any parameter ...)  If the floating point unit
109
// is not included and OPT_ILLEGAL_INSTRUCTION is set, then as with the
110
// multiply and divide any floating point instruction will result in an illegal
111
// instruction exception that will send the CPU into supervisor mode.
112
//
113
//
114
// `define      OPT_IMPLEMENT_FPU
115
//
116
//
117
//
118
// OPT_NEW_INSTRUCTION_SET controls whether or not the new instruction set
119
// is in use.  The new instruction set contains space for floating point
120
// operations, signed and unsigned divide instructions, as well as bit reversal
121
// and ... at least two other operations yet to be defined.  The decoder alone
122
// uses about 70 fewer LUTs, although in practice this works out to 12 fewer
123
// when all works out in the wash.  Further, floating point and divide
124
// instructions will cause an illegal instruction exception if they are not
125
// implemented--so software capability can be built to use these instructions
126
// immediately, even if the hardware is not yet ready.
127
//
128
// This option is likely to go away in the future, obsoleting the previous
129
// instruction set, so I recommend setting this option and switching to the
130
// new instruction set as soon as possible.
131
//
132
`define OPT_NEW_INSTRUCTION_SET
133
//
134
//
135
//
136
//
137
//
138
//
139
// OPT_SINGLE_FETCH controls whether or not the prefetch has a cache, and 
140
// whether or not it can issue one instruction per clock.  When set, the
141
// prefetch has no cache, and only one instruction is fetched at a time.
142
// This effectively sets the CPU so that only one instruction is ever 
143
// in the pipeline at once, and hence you may think of this as a "kill 
144
// pipeline" option.  However, since the pipelined fetch component uses so
145
// much area on the FPGA, this is an important option to use in trimming down
146
// used area if necessary.  Hence, it needs to be maintained for that purpose.
147
// Be aware, though, it will drop your performance by a factor between 2x and
148
// 3x.
149
//
150
// We can either pipeline our fetches, or issue one fetch at a time.  Pipelined
151
// fetches are more complicated and therefore use more FPGA resources, while
152
// single fetches will cause the CPU to stall for about 5 stalls each 
153
// instruction cycle, effectively reducing the instruction count per clock to
154
// about 0.2.  However, the area cost may be worth it.  Consider:
155
//
156
//      Slice LUTs              ZipSystem       ZipCPU
157
//      Single Fetching         2521            1734
158
//      Pipelined fetching      2796            2046
159
//      (These numbers may be dated, but should still be representative ...)
160
//
161
// I recommend only defining this if you "need" to, if area is tight and
162
// speed isn't as important.  Otherwise, just leave this undefined.
163
//
164
// `define      OPT_SINGLE_FETCH
165
//
166
//
167
//
168
// The next several options are pipeline optimization options.  They make no
169
// sense in a single instruction fetch mode, hence we #ifndef them so they
170
// are only defined if we are in a full pipelined mode (i.e. OPT_SINGLE_FETCH
171
// is not defined).
172
//
173
`ifndef OPT_SINGLE_FETCH
174
//
175
//
176
//
177
// OPT_PIPELINED is the natural result and opposite of using the single 
178
// instruction fetch unit.  If you are not using that unit, the ZipCPU will
179
// be pipelined.  The option is defined here more for readability than 
180
// anything else, since OPT_PIPELINED makes more sense than OPT_SINGLE_FETCH,
181
// well ... that and it does a better job of explaining what is going on.
182
//
183
// In other words, leave this define alone--lest you break the ZipCPU.
184
//
185
`define OPT_PIPELINED
186
//
187
//
188
//
189
// OPT_TRADITIONAL_PFCACHE allows you to switch between one of two prefetch
190
// caches.  If enabled, a more traditional cache is implemented.  This more
191
// traditional cache (currently) uses many more LUTs, but it also reduces
192
// the stall count tremendously over the alternative hacked pipeline cache.
193
// (The traditional pfcache is also pipelined, whereas the pipeline cache
194
// implements a windowed approach to caching.)
195
//
196
// If you have the fabric to support this option, I recommend including it.
197
//
198
`define OPT_TRADITIONAL_PFCACHE
199
//
200
//
201
//
202
// OPT_EARLY_BRANCHING is an attempt to execute a BRA statement as early
203
// as possible, to avoid as many pipeline stalls on a branch as possible.
204
// It's not tremendously successful yet--BRA's still suffer stalls,
205
// but I intend to keep working on this approach until the number of stalls
206
// gets down to one or (ideally) zero.  (With the OPT_TRADITIONAL_PFCACHE, this
207
// gets down to a single stall cycle ...)  That way a "BRA" can be used as the
208
// compiler's branch prediction optimizer: BRA's barely stall, while branches
209
// on conditions will always suffer about 4 stall cycles or so.
210
//
211
// I recommend setting this flag, so as to turn early branching on.
212
//
213
`define OPT_EARLY_BRANCHING
214
//
215
//
216
//
217
// OPT_PIPELINED_BUS_ACCESS controls whether or not LOD/STO instructions
218
// can take advantaged of pipelined bus instructions.  To be eligible, the
219
// operations must be identical (cannot pipeline loads and stores, just loads
220
// only or stores only), and the addresses must either be identical or one up
221
// from the previous address.  Further, the load/store string must all have
222
// the same conditional.  This approach gains the must use, in my humble
223
// opinion, when saving registers to or restoring registers from the stack
224
// at the beginning/end of a procedure, or when doing a context swap.
225
//
226
// I recommend setting this flag, for performance reasons, especially if your
227
// wishbone bus can handle pipelined bus accesses.
228
//
229
`define OPT_PIPELINED_BUS_ACCESS
230
//
231
//
232
//
233
`ifdef  OPT_NEW_INSTRUCTION_SET
234
//
235
//
236
//
237
// The new instruction set also defines a set of very long instruction words.
238
// Well, calling them "very long" instruction words is probably a misnomer,
239
// although we're going to do it.  They're really 2x16-bit instructions---
240
// instruction words that pack two instructions into one word.  (2x14 bit
241
// really--'cause you need a bit to note the instruction is a 2x instruction,
242
// and then 3-bits for the condition codes ...)  Set OPT_VLIW to include these
243
// double instructions as part of the new instruction set.  These allow a single
244
// instruction to contain two instructions within.   These instructions are
245
// designed to get more code density from the instruction set, and to hopefully
246
// take some pain off of the performance of the pre-fetch and instruction cache.
247
//
248
// These new instructions, however, also necessitate a change in the Zip
249
// CPU--the Zip CPU can no longer execute instructions atomically.  It must
250
// now execute non-VLIW instructions, or VLIW instruction pairs, atomically. 
251
// This logic has been added into the ZipCPU, but it has not (yet) been
252
// tested thoroughly.
253
//
254
// Oh, and the assembler, the debugger, and the object file dumper, and the
255
// simulator all need to be updated as well ....
256
//
257
`define OPT_VLIW
258
//
259
//
260
`endif // OPT_NEW_INSTRUCTION_SET
261
//
262
//
263
`endif  // OPT_SINGLE_FETCH
264
//
265
//
266
//
267
// Now let's talk about peripherals for a moment.  These next two defines
268
// control whether the DMA controller is included in the Zip System, and
269
// whether or not the 8 accounting timers are also included.  Set these to
270
// include the respective peripherals, comment them out not to.
271
//
272
`define INCLUDE_DMA_CONTROLLER
273
`define INCLUDE_ACCOUNTING_COUNTERS
274
//
275
//
276 30 dgisselq
`define DEBUG_SCOPE
277 3 dgisselq
//
278
`endif  // CPUDEFS_H

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.