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
|
// =================================================================================================
// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. This
// project loosely follows the Google C++ styleguide and uses a tab-size of two spaces and a max-
// width of 100 characters per line.
//
// Author(s):
// Cedric Nugteren <www.cedricnugteren.nl>
//
// This file implements a class with static methods to describe the Xcol2im routine. Examples of
// such 'descriptions' are how to calculate the size a of buffer or how to run the routine. These
// static methods are used by the correctness tester and the performance tester.
//
// =================================================================================================
#ifndef CLBLAST_TEST_ROUTINES_XCOL2IM_H_
#define CLBLAST_TEST_ROUTINES_XCOL2IM_H_
#include "test/routines/common.hpp"
namespace clblast {
// =================================================================================================
// See comment at top of file for a description of the class
template <typename T>
class TestXcol2im {
public:
// The BLAS level: 4 for the extra routines
static size_t BLASLevel() { return 4; }
// The list of arguments relevant for this routine
static std::vector<std::string> GetOptions() {
return {kArgKernelMode,
kArgChannels, kArgHeight, kArgWidth, kArgKernelH, kArgKernelW, kArgPadH, kArgPadW,
kArgStrideH, kArgStrideW, kArgDilationH, kArgDilationW,
kArgAOffset, kArgBOffset};
}
static std::vector<std::string> BuffersIn() { return {kBufMatA, kBufMatB}; } // b = col
static std::vector<std::string> BuffersOut() { return {kBufMatA}; } // a = im
// Describes how to obtain the sizes of the buffers
static size_t ColHeight(const Arguments<T> &args) {
const auto size = args.height + 2 * args.pad_h;
const auto padding = args.dilation_h * (args.kernel_h - 1) + 1;
if (size >= padding) { return (size - padding) / args.stride_h + 1; }
return 1;
}
static size_t ColWidth(const Arguments<T> &args) {
const auto size = args.width + 2 * args.pad_w;
const auto padding = args.dilation_w * (args.kernel_w - 1) + 1;
if (size >= padding) { return (size - padding) / args.stride_w + 1; }
return 1;
}
static size_t NumPatches(const Arguments<T> &args) {
return ColHeight(args) * ColWidth(args) * args.channels;
}
static size_t GetSizeA(const Arguments<T> &args) {
return args.height * args.width * args.channels + args.a_offset;
}
static size_t GetSizeB(const Arguments<T> &args) {
return args.kernel_w * args.kernel_h * NumPatches(args) + args.b_offset;
}
// Describes how to set the sizes of all the buffers
static void SetSizes(Arguments<T> &args, Queue&) {
args.a_size = GetSizeA(args); // im
args.b_size = GetSizeB(args); // col
}
// Describes what the default values of the leading dimensions of the matrices are
static size_t DefaultLDA(const Arguments<T> &) { return 1; } // N/A for this routine
static size_t DefaultLDB(const Arguments<T> &) { return 1; } // N/A for this routine
static size_t DefaultLDC(const Arguments<T> &) { return 1; } // N/A for this routine
// Describes which transpose options are relevant for this routine
using Transposes = std::vector<Transpose>;
static Transposes GetATransposes(const Transposes &) { return {}; } // N/A for this routine
static Transposes GetBTransposes(const Transposes &) { return {}; } // N/A for this routine
// Describes how to prepare the input data
static void PrepareData(const Arguments<T>&, Queue&, const int, std::vector<T>&,
std::vector<T>&, std::vector<T>&, std::vector<T>&, std::vector<T>&,
std::vector<T>&, std::vector<T>&) {} // N/A for this routine
// Describes how to run the CLBlast routine
static StatusCode RunRoutine(const Arguments<T> &args, Buffers<T> &buffers, Queue &queue) {
#ifdef OPENCL_API
auto queue_plain = queue();
auto event = cl_event{};
auto status = Col2im<T>(args.kernel_mode,
args.channels, args.height, args.width,
args.kernel_h, args.kernel_w,
args.pad_h, args.pad_w,
args.stride_h, args.stride_w,
args.dilation_h, args.dilation_w,
buffers.b_mat(), args.b_offset, // col
buffers.a_mat(), args.a_offset, // im
&queue_plain, &event);
if (status == StatusCode::kSuccess) { clWaitForEvents(1, &event); clReleaseEvent(event); }
#elif CUDA_API
auto status = Col2im<T>(args.kernel_mode,
args.channels, args.height, args.width,
args.kernel_h, args.kernel_w,
args.pad_h, args.pad_w,
args.stride_h, args.stride_w,
args.dilation_h, args.dilation_w,
buffers.b_mat(), args.b_offset, // col
buffers.a_mat(), args.a_offset, // im
queue.GetContext()(), queue.GetDevice()());
cuStreamSynchronize(queue());
#endif
return status;
}
// Describes how to run a naive version of the routine (for correctness/performance comparison).
// Note that a proper clBLAS or CPU BLAS comparison is not available for non-BLAS routines.
static StatusCode RunReference1(const Arguments<T> &args, Buffers<T> &buffers, Queue &queue) {
auto buffers_host = BuffersHost<T>();
DeviceToHost(args, buffers, buffers_host, queue, BuffersIn());
const auto status = RunReference(args, buffers_host);
HostToDevice(args, buffers, buffers_host, queue, BuffersOut());
return status;
}
static StatusCode RunReference2(const Arguments<T> &args, BuffersHost<T> &buffers_host, Queue&) {
return RunReference(args, buffers_host);
}
static StatusCode RunReference3(const Arguments<T> &, BuffersCUDA<T> &, Queue &) {
return StatusCode::kUnknownError;
}
// Describes how to download the results of the computation (more importantly: which buffer)
static std::vector<T> DownloadResult(const Arguments<T> &args, Buffers<T> &buffers, Queue &queue) {
std::vector<T> result(args.a_size, static_cast<T>(0));
buffers.a_mat.Read(queue, args.a_size, result);
return result;
}
// Describes how to compute the indices of the result buffer
static size_t ResultID1(const Arguments<T> &args) { return args.height * args.width; }
static size_t ResultID2(const Arguments<T> &args) { return args.channels; }
static size_t GetResultIndex(const Arguments<T> &args, const size_t id1, const size_t id2) {
return id1 + args.height * args.width * id2 + args.a_offset;
}
// Describes how to compute performance metrics
static size_t GetFlops(const Arguments<T> &) {
return 1;
}
static size_t GetBytes(const Arguments<T> &args) {
const auto im = args.channels * args.width * args.height; // possibly less with striding
const auto col = args.kernel_h * args.kernel_w * NumPatches(args);
return (im + col) * sizeof(T);
}
};
// =================================================================================================
template <typename T>
StatusCode RunReference(const Arguments<T> &args, BuffersHost<T> &buffers_host) {
// Reference taken from im2col but swapped the input/output
const auto col_h = TestXcol2im<T>::ColHeight(args);
const auto col_w = TestXcol2im<T>::ColWidth(args);
for (auto c_id = size_t{0}; c_id < args.channels; ++c_id) { // image channels
for (auto kh_id = size_t{0}; kh_id < args.kernel_h; ++kh_id) { // kernel height
for (auto kw_id = size_t{0}; kw_id < args.kernel_w; ++kw_id) { // kernel width
for (auto h_id = size_t{0}; h_id < col_h; ++h_id) { // image height
for (auto w_id = size_t{0}; w_id < col_w; ++w_id) { // image width
// Reads the input value
const auto kernel_index
= (args.kernel_mode == KernelMode::kConvolution)
? args.kernel_h * args.kernel_w - kw_id - args.kernel_w * kh_id - 1
: kw_id + args.kernel_w * kh_id;
const auto patch_index = w_id + col_w * h_id;
const auto col_index = patch_index + kernel_index * col_w * col_h +
c_id * col_w * col_h * args.kernel_h * args.kernel_w;
const auto val = buffers_host.b_mat[col_index + args.b_offset];
// Sets the output value
const auto h_index = kh_id * args.dilation_h + args.stride_h * h_id - args.pad_h;
const auto w_index = kw_id * args.dilation_w + args.stride_w * w_id - args.pad_w;
if (h_index >= 0 && h_index < args.height &&
w_index >= 0 && w_index < args.width) {
const auto im_index = w_index + args.width * (h_index + args.height * c_id);
buffers_host.a_mat[im_index + args.a_offset] += val;
}
}
}
}
}
}
return StatusCode::kSuccess;
}
// Half-precision version calling the above reference implementation after conversions
template <>
StatusCode RunReference<half>(const Arguments<half> &args, BuffersHost<half> &buffers_host) {
auto a_buffer2 = HalfToFloatBuffer(buffers_host.a_mat);
auto b_buffer2 = HalfToFloatBuffer(buffers_host.b_mat);
auto dummy = std::vector<float>(0);
auto buffers2 = BuffersHost<float>{dummy, dummy, a_buffer2, b_buffer2, dummy, dummy, dummy};
auto args2 = Arguments<float>();
args2.a_size = args.a_size; args2.b_size = args.b_size;
args2.kernel_mode = args.kernel_mode;
args2.channels = args.channels; args2.height = args.height; args2.width = args.width;
args2.kernel_h = args.kernel_h; args2.kernel_w = args.kernel_w;
args2.pad_h = args.pad_h; args2.pad_w = args.pad_w;
args2.stride_h = args.stride_h; args2.stride_w = args.stride_w;
args2.dilation_h = args.dilation_h; args2.dilation_w = args.dilation_w;
args2.a_offset = args.a_offset; args2.b_offset = args.b_offset;
auto status = RunReference(args2, buffers2);
FloatToHalfBuffer(buffers_host.a_mat, buffers2.a_mat);
return status;
}
// =================================================================================================
} // namespace clblast
// CLBLAST_TEST_ROUTINES_XCOL2IM_H_
#endif
|
