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
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
|
/*
Copyright (C) 1999-2008 by Mark D. Hill and David A. Wood for the
Wisconsin Multifacet Project. Contact: gems@cs.wisc.edu
http://www.cs.wisc.edu/gems/
--------------------------------------------------------------------
This file is part of the Ruby Multiprocessor Memory System Simulator,
a component of the Multifacet GEMS (General Execution-driven
Multiprocessor Simulator) software toolset originally developed at
the University of Wisconsin-Madison.
Ruby was originally developed primarily by Milo Martin and Daniel
Sorin with contributions from Ross Dickson, Carl Mauer, and Manoj
Plakal.
Substantial further development of Multifacet GEMS at the
University of Wisconsin was performed by Alaa Alameldeen, Brad
Beckmann, Jayaram Bobba, Ross Dickson, Dan Gibson, Pacia Harper,
Derek Hower, Milo Martin, Michael Marty, Carl Mauer, Michelle Moravan,
Kevin Moore, Andrew Phelps, Manoj Plakal, Daniel Sorin, Haris Volos,
Min Xu, and Luke Yen.
--------------------------------------------------------------------
If your use of this software contributes to a published paper, we
request that you (1) cite our summary paper that appears on our
website (http://www.cs.wisc.edu/gems/) and (2) e-mail a citation
for your published paper to gems@cs.wisc.edu.
If you redistribute derivatives of this software, we request that
you notify us and either (1) ask people to register with us at our
website (http://www.cs.wisc.edu/gems/) or (2) collect registration
information and periodically send it to us.
--------------------------------------------------------------------
Multifacet GEMS is free software; you can redistribute it and/or
modify it under the terms of version 2 of the GNU General Public
License as published by the Free Software Foundation.
Multifacet GEMS is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.
You should have received a copy of the GNU General Public License
along with the Multifacet GEMS; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
02111-1307, USA
The GNU General Public License is contained in the file LICENSE.
### END HEADER ###
*/
/*
* $Id$
*
*/
#include "mem/ruby/common/Global.hh"
#include "mem/ruby/tester/Tester_Globals.hh"
#include "mem/ruby/tester/DeterministicDriver.hh"
#include "mem/ruby/eventqueue/RubyEventQueue.hh"
//#include "DMAGenerator.hh"
#include "mem/ruby/tester/DetermGETXGenerator.hh"
#define DATA_BLOCK_BYTES 64
DeterministicDriver::DeterministicDriver(string generator_type, int num_completions, int num_procs, Time g_think_time, Time g_wait_time, int g_tester_length)
{
eventQueue = new RubyEventQueue;
m_finish_time = 0;
m_last_issue = -11;
m_done_counter = 0;
m_loads_completed = 0;
m_stores_completed = 0;
m_numCompletionsPerNode = num_completions;
m_num_procs = num_procs;
m_think_time = g_think_time;
m_wait_time = g_wait_time;
m_tester_length = g_tester_length;
m_last_progress_vector.setSize(num_procs);
for (int i=0; i<m_last_progress_vector.size(); i++) {
m_last_progress_vector[i] = 0;
}
m_load_vector.setSize(10);
for (int i=0; i<m_load_vector.size(); i++) {
m_load_vector[i] = -1; // No processor last held it
}
m_store_vector.setSize(10);
for (int i=0; i<m_store_vector.size(); i++) {
m_store_vector[i] = -1; // No processor last held it
}
m_generator_vector.setSize(num_procs);
int generator = string_to_SpecifiedGeneratorType(generator_type);
for (int i=0; i<m_generator_vector.size(); i++) {
switch (generator) {
case SpecifiedGeneratorType_DetermGETXGenerator:
m_generator_vector[i] = new DetermGETXGenerator(i, this);
break;
case SpecifiedGeneratorType_DetermInvGenerator:
m_generator_vector[i] = new DetermInvGenerator(i, *this);
break;
case SpecifiedGeneratorType_DetermSeriesGETSGenerator:
m_generator_vector[i] = new DetermSeriesGETSGenerator(i, *this);
break;
default:
ERROR_MSG("Unexpected specified generator type");
}
}
//m_dma_generator = new DMAGenerator(0, this);
}
void DeterministicDriver::go()
{
// tick both queues until everyone is done
while (m_done_counter != m_num_procs) {
libruby_tick(1);
eventQueue->triggerEvents(eventQueue->getTime() + 1);
}
}
DeterministicDriver::~DeterministicDriver()
{
for (int i=0; i<m_last_progress_vector.size(); i++) {
delete m_generator_vector[i];
}
}
//void DeterministicDriver::dmaHitCallback()
//{
// m_dma_generator->performCallback();
//}
void DeterministicDriver::wakeup() {
assert(0);
// this shouldn't be called as we are not scheduling the driver ever
}
void DeterministicDriver::hitCallback(int64_t request_id)
{
ASSERT(requests.find(request_id) != requests.end());
int proc = requests[request_id].first;
Address address = requests[request_id].second;
m_generator_vector[proc]->performCallback(proc, address);
m_last_progress_vector[proc] = eventQueue->getTime();
requests.erase(request_id);
}
bool DeterministicDriver::isStoreReady(NodeID node)
{
return isAddrReady(node, m_store_vector);
}
bool DeterministicDriver::isStoreReady(NodeID node, Address addr)
{
int addr_number = addr.getAddress()/DATA_BLOCK_BYTES;
return isAddrReady(node, m_store_vector, addr);
}
bool DeterministicDriver::isLoadReady(NodeID node)
{
return isAddrReady(node, m_load_vector);
}
bool DeterministicDriver::isLoadReady(NodeID node, Address addr)
{
return isAddrReady(node, m_load_vector, addr);
}
// searches for any address in the addr_vector
bool DeterministicDriver::isAddrReady(NodeID node, Vector<NodeID> addr_vector)
{
for (int i=0; i<addr_vector.size(); i++) {
if (((addr_vector[i]+1)%m_num_procs == node) &&
(m_loads_completed+m_stores_completed >= m_numCompletionsPerNode*node) && // is this node next
(eventQueue->getTime() >= m_last_issue + 10)) { // controll rate of requests
return true;
}
}
return false;
}
// test for a particular addr
bool DeterministicDriver::isAddrReady(NodeID node, Vector<NodeID> addr_vector, Address addr)
{
int addr_number = addr.getAddress()/DATA_BLOCK_BYTES;
ASSERT ((addr_number >= 0) && (addr_number < addr_vector.size()));
if (((addr_vector[addr_number]+1)%m_num_procs == node) &&
(m_loads_completed+m_stores_completed >= m_numCompletionsPerNode*node) && // is this node next
(eventQueue->getTime() >= m_last_issue + 10)) { // controll rate of requests
return true;
} else {
return false;
}
}
void DeterministicDriver::loadCompleted(NodeID node, Address addr)
{
m_loads_completed++;
setNextAddr(node, addr, m_load_vector);
}
void DeterministicDriver::storeCompleted(NodeID node, Address addr)
{
m_stores_completed++;
setNextAddr(node, addr, m_store_vector);
}
void DeterministicDriver::setNextAddr(NodeID node, Address addr, Vector<NodeID>& addr_vector)
{
// mark the addr vector that this proc was the last to use the particular address
int addr_number = addr.getAddress()/DATA_BLOCK_BYTES;
addr_vector[addr_number] = node;
}
Address DeterministicDriver::getNextLoadAddr(NodeID node)
{
return getNextAddr(node, m_load_vector);
}
Address DeterministicDriver::getNextStoreAddr(NodeID node)
{
return getNextAddr(node, m_store_vector);
}
Address DeterministicDriver::getNextAddr(NodeID node, Vector<NodeID> addr_vector)
{
// This method deterministically picks the next addr the node should acquirer
// The addrs cycle through according to NodeID 0->1->...->lastID->0...
Address addr;
// should only be called if we know a addr is ready for the node
ASSERT(isAddrReady(node, addr_vector));
for (int addr_number=0; addr_number<addr_vector.size(); addr_number++) {
// is this node next in line for the addr
// POLINA: LOOK HERE!
// if ((addr_vector[addr_number] != 1) && ((addr_vector[addr_number]+1)%m_num_procs) == node) {
if (((addr_vector[addr_number]+1)%m_num_procs) == node) {
// One addr per cache line
addr.setAddress(addr_number * DATA_BLOCK_BYTES);
}
}
m_last_issue = eventQueue->getTime();
return addr;
}
void DeterministicDriver::reportDone()
{
m_done_counter++;
if ((m_done_counter == m_num_procs)) {
m_finish_time = eventQueue->getTime();
//m_dma_generator->stop();
}
}
void DeterministicDriver::recordLoadLatency(Time time)
{
m_load_latency.add(time);
}
void DeterministicDriver::recordStoreLatency(Time time)
{
m_store_latency.add(time);
}
void DeterministicDriver::printStats(ostream& out) const
{
out << endl;
out << "DeterministicDriver Stats" << endl;
out << "---------------------" << endl;
out << "finish_time: " << m_finish_time << endl;
out << "load_latency: " << m_load_latency << endl;
out << "store_latency: " << m_store_latency << endl;
}
void DeterministicDriver::print(ostream& out) const
{
}
|