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
|
/*
* l2cache.h
*
* Copyright (c) 2009 by Tor M. Aamodt and
* University of British Columbia
* Vancouver, BC V6T 1Z4
* All Rights Reserved.
*
* THIS IS A LEGAL DOCUMENT BY DOWNLOADING GPGPU-SIM, YOU ARE AGREEING TO THESE
* TERMS AND CONDITIONS.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNERS OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*
* NOTE: The files libcuda/cuda_runtime_api.c and src/cuda-sim/cuda-math.h
* are derived from the CUDA Toolset available from http://www.nvidia.com/cuda
* (property of NVIDIA). The files benchmarks/BlackScholes/ and
* benchmarks/template/ are derived from the CUDA SDK available from
* http://www.nvidia.com/cuda (also property of NVIDIA). The files from
* src/intersim/ are derived from Booksim (a simulator provided with the
* textbook "Principles and Practices of Interconnection Networks" available
* from http://cva.stanford.edu/books/ppin/). As such, those files are bound by
* the corresponding legal terms and conditions set forth separately (original
* copyright notices are left in files from these sources and where we have
* modified a file our copyright notice appears before the original copyright
* notice).
*
* Using this version of GPGPU-Sim requires a complete installation of CUDA
* which is distributed seperately by NVIDIA under separate terms and
* conditions. To use this version of GPGPU-Sim with OpenCL requires a
* recent version of NVIDIA's drivers which support OpenCL.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* 3. Neither the name of the University of British Columbia nor the names of
* its contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* 4. This version of GPGPU-SIM is distributed freely for non-commercial use only.
*
* 5. No nonprofit user may place any restrictions on the use of this software,
* including as modified by the user, by any other authorized user.
*
* 6. GPGPU-SIM was developed primarily by Tor M. Aamodt, Wilson W. L. Fung,
* Ali Bakhoda, George L. Yuan, at the University of British Columbia,
* Vancouver, BC V6T 1Z4
*/
#ifndef MC_PARTITION_INCLUDED
#define MC_PARTITION_INCLUDED
#include "dram.h"
#include "../abstract_hardware_model.h"
#include <list>
#include <queue>
class mem_fetch;
class memory_partition_unit
{
public:
memory_partition_unit( unsigned partition_id, const struct memory_config *config, class memory_stats_t *stats );
~memory_partition_unit();
bool busy() const;
void cache_cycle( unsigned cycle );
void dram_cycle();
bool full() const;
void push( class mem_fetch* mf, unsigned long long clock_cycle );
class mem_fetch* pop();
class mem_fetch* top();
unsigned flushL2();
void visualizer_print( gzFile visualizer_file );
void print_cache_stat(unsigned &accesses, unsigned &misses) const;
void print_stat( FILE *fp ) { m_dram->print_stat(fp); }
void visualize() const { m_dram->visualize(); }
void print( FILE *fp ) const;
private:
// data
unsigned m_id;
const struct memory_config *m_config;
class dram_t *m_dram;
class read_only_cache *m_L2cache;
class L2interface *m_L2interface;
// model delay of ROP units with a fixed latency
struct rop_delay_t
{
unsigned long long ready_cycle;
class mem_fetch* req;
};
std::queue<rop_delay_t> m_rop;
// these are various FIFOs between units within a memory partition
fifo_pipeline<mem_fetch> *m_icnt_L2_queue;
fifo_pipeline<mem_fetch> *m_L2_dram_queue;
fifo_pipeline<mem_fetch> *m_dram_L2_queue;
fifo_pipeline<mem_fetch> *m_L2_icnt_queue; // L2 cache hit response queue
class mem_fetch *L2dramout;
unsigned long long int wb_addr;
class memory_stats_t *m_stats;
std::set<mem_fetch*> m_request_tracker;
friend class L2interface;
};
class L2interface : public mem_fetch_interface {
public:
L2interface( memory_partition_unit *unit ) { m_unit=unit; }
virtual bool full( unsigned size, bool write) const
{
// assume read and write packets all same size
return m_unit->m_L2_dram_queue->full();
}
virtual void push(mem_fetch *mf)
{
m_unit->m_L2_dram_queue->push(mf);
}
private:
memory_partition_unit *m_unit;
};
#endif
|