forked from NVIDIA/cuda-samples
-
Notifications
You must be signed in to change notification settings - Fork 0
/
sharedmem.cuh
173 lines (157 loc) · 4.89 KB
/
sharedmem.cuh
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
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * 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.
* * Neither the name of NVIDIA CORPORATION nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``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 OWNER 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.
*/
#ifndef _SHAREDMEM_H_
#define _SHAREDMEM_H_
//****************************************************************************
// Because dynamically sized shared memory arrays are declared "extern",
// we can't templatize them directly. To get around this, we declare a
// simple wrapper struct that will declare the extern array with a different
// name depending on the type. This avoids compiler errors about duplicate
// definitions.
//
// To use dynamically allocated shared memory in a templatized __global__ or
// __device__ function, just replace code like this:
//
//
// template<class T>
// __global__ void
// foo( T* g_idata, T* g_odata)
// {
// // Shared mem size is determined by the host app at run time
// extern __shared__ T sdata[];
// ...
// doStuff(sdata);
// ...
// }
//
// With this
// template<class T>
// __global__ void
// foo( T* g_idata, T* g_odata)
// {
// // Shared mem size is determined by the host app at run time
// SharedMemory<T> smem;
// T* sdata = smem.getPointer();
// ...
// doStuff(sdata);
// ...
// }
//****************************************************************************
// This is the un-specialized struct. Note that we prevent instantiation of
// this
// struct by putting an undefined symbol in the function body so it won't
// compile.
template <typename T>
struct SharedMemory {
// Ensure that we won't compile any un-specialized types
__device__ T *getPointer() {
extern __device__ void error(void);
error();
return NULL;
}
};
// Following are the specializations for the following types.
// int, uint, char, uchar, short, ushort, long, ulong, bool, float, and double
// One could also specialize it for user-defined types.
template <>
struct SharedMemory<int> {
__device__ int *getPointer() {
extern __shared__ int s_int[];
return s_int;
}
};
template <>
struct SharedMemory<unsigned int> {
__device__ unsigned int *getPointer() {
extern __shared__ unsigned int s_uint[];
return s_uint;
}
};
template <>
struct SharedMemory<char> {
__device__ char *getPointer() {
extern __shared__ char s_char[];
return s_char;
}
};
template <>
struct SharedMemory<unsigned char> {
__device__ unsigned char *getPointer() {
extern __shared__ unsigned char s_uchar[];
return s_uchar;
}
};
template <>
struct SharedMemory<short> {
__device__ short *getPointer() {
extern __shared__ short s_short[];
return s_short;
}
};
template <>
struct SharedMemory<unsigned short> {
__device__ unsigned short *getPointer() {
extern __shared__ unsigned short s_ushort[];
return s_ushort;
}
};
template <>
struct SharedMemory<long> {
__device__ long *getPointer() {
extern __shared__ long s_long[];
return s_long;
}
};
template <>
struct SharedMemory<unsigned long> {
__device__ unsigned long *getPointer() {
extern __shared__ unsigned long s_ulong[];
return s_ulong;
}
};
template <>
struct SharedMemory<bool> {
__device__ bool *getPointer() {
extern __shared__ bool s_bool[];
return s_bool;
}
};
template <>
struct SharedMemory<float> {
__device__ float *getPointer() {
extern __shared__ float s_float[];
return s_float;
}
};
template <>
struct SharedMemory<double> {
__device__ double *getPointer() {
extern __shared__ double s_double[];
return s_double;
}
};
#endif //_SHAREDMEM_H_