-
Notifications
You must be signed in to change notification settings - Fork 0
/
cudaMemPrefetchAsync.cu
204 lines (160 loc) · 8.4 KB
/
cudaMemPrefetchAsync.cu
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
/*
* Copyright 1993-2020 NVIDIA Corporation. All rights reserved.
*
* NOTICE TO USER:
*
* This source code is subject to NVIDIA ownership rights under U.S. and
* international Copyright laws. Users and possessors of this source code
* are hereby granted a nonexclusive, royalty-free license to use this code
* in individual and commercial software.
*
* NVIDIA MAKES NO REPRESENTATION ABOUT THE SUITABILITY OF THIS SOURCE
* CODE FOR ANY PURPOSE. IT IS PROVIDED "AS IS" WITHOUT EXPRESS OR
* IMPLIED WARRANTY OF ANY KIND. NVIDIA DISCLAIMS ALL WARRANTIES WITH
* REGARD TO THIS SOURCE CODE, INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY, NONINFRINGEMENT, AND FITNESS FOR A PARTICULAR PURPOSE.
* IN NO EVENT SHALL NVIDIA BE LIABLE FOR ANY SPECIAL, INDIRECT, INCIDENTAL,
* OR CONSEQUENTIAL DAMAGES, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS
* OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
* OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE
* OR PERFORMANCE OF THIS SOURCE CODE.
*
* U.S. Government End Users. This source code is a "commercial item" as
* that term is defined at 48 C.F.R. 2.101 (OCT 1995), consisting of
* "commercial computer software" and "commercial computer software
* documentation" as such terms are used in 48 C.F.R. 12.212 (SEPT 1995)
* and is provided to the U.S. Government only as a commercial end item.
* Consistent with 48 C.F.R.12.212 and 48 C.F.R. 227.7202-1 through
* 227.7202-4 (JUNE 1995), all U.S. Government End Users acquire the
* source code with only those rights set forth herein.
*
* Any use of this source code in individual and commercial software must
* include, in the user documentation and internal comments to the code,
* the above Disclaimer and U.S. Government End Users Notice.
*/
/*
* This example benchmarks copying two data arrays to and from the GPU.
* It uses unified memory memory and chrono, cuda events, and NVTX for timing.
*/
#include "cuda_helper.h"
template<typename T>
void MemManaged<T>::run_chrono( ) {
for ( int i = 0; i < this->loops; i++ ) {
this->reset( N, h_a_unified.get( ), h_b_unified.get( ) );
this->start = std::chrono::high_resolution_clock::now( );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, this->device, this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, this->device, this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Add<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->a_args,
0,
this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Sub<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->b_args,
0,
this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[1] ) );
this->verify( N, h_a_unified.get( ), h_b_unified.get( ) );
this->stop = std::chrono::high_resolution_clock::now( );
this->elapsed_chrono_ms += this->stop - this->start;
}
this->get_chrono_results( size );
}
template<typename T>
void MemManaged<T>::run_events( ) {
for ( int i = 0; i < this->loops; i++ ) {
this->reset( N, h_a_unified.get( ), h_b_unified.get( ) );
CUDA_RT_CALL( cudaEventRecord( this->start_event ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, this->device, this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, this->device, this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Add<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->a_args,
0,
this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Sub<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->b_args,
0,
this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[1] ) );
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[1] ) );
this->verify( N, h_a_unified.get( ), h_b_unified.get( ) );
CUDA_RT_CALL( cudaEventRecord( this->stop_event ) );
CUDA_RT_CALL( cudaEventSynchronize( this->stop_event ) );
CUDA_RT_CALL( cudaEventElapsedTime( &this->elapsed_events_ms, this->start_event, this->stop_event ) );
this->average_events_ms += this->elapsed_events_ms;
}
this->get_events_results( size );
}
template<typename T>
void MemManaged<T>::run_nvtx( ) {
for ( int i = 0; i < this->loops; i++ ) {
PUSH_RANGE( "Reset", 0 )
this->reset( N, h_a_unified.get( ), h_b_unified.get( ) );
POP_RANGE( )
PUSH_RANGE( "Process_Loop", 1 )
PUSH_RANGE( "H2D_A", 2 )
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, this->device, this->cuda_streams[0] ) );
POP_RANGE( )
PUSH_RANGE( "H2D_B", 3 )
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, this->device, this->cuda_streams[1] ) );
POP_RANGE( )
PUSH_RANGE( "Kernel_A", 4 )
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Add<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->a_args,
0,
this->cuda_streams[0] ) );
POP_RANGE( )
PUSH_RANGE( "Kernel_B", 5 )
CUDA_RT_CALL( cudaLaunchKernel( reinterpret_cast<void *>( &VectorOperation<Sub<T>, T> ),
this->blocks_per_grid,
this->threads_per_block,
this->b_args,
0,
this->cuda_streams[1] ) );
POP_RANGE( )
PUSH_RANGE( "D2H_A", 6 )
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_a_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[0] ) );
POP_RANGE( )
PUSH_RANGE( "D2H_B", 7 )
CUDA_RT_CALL( cudaMemPrefetchAsync( this->h_b_unified.get( ), size, cudaCpuDeviceId, this->cuda_streams[1] ) );
POP_RANGE( )
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[0] ) );
CUDA_RT_CALL( cudaStreamSynchronize( this->cuda_streams[1] ) );
PUSH_RANGE( "Verify", 8 )
this->verify( N, h_a_unified.get( ), h_b_unified.get( ) );
POP_RANGE( )
POP_RANGE( )
}
}
/* Main */
int main( int argc, char **argv ) {
using dtype = float;
int N = MemCpy<dtype>::init_size;
if ( argc > 1 ) {
N = std::atoi( argv[1] );
}
MemManaged<dtype> MemManaged( N );
double gigabytes { MemManaged.size * 1e-9 };
printf( "Running with = %lu B (%0.2f GB)\n\n", MemManaged.size, gigabytes );
// Chrono
MemManaged.run_chrono( );
// Events
MemManaged.run_events( );
// NVTX
MemManaged.run_nvtx( );
return ( EXIT_SUCCESS );
}