node-native-win-utils 1.4.0 → 2.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +29 -82
- package/binding.gyp +2 -0
- package/dist/dirnameLocal.cjs +4 -0
- package/dist/dirnameLocal.d.cts +1 -0
- package/dist/dirnameLocal.d.mts +1 -0
- package/dist/dirnameLocal.mjs +2 -0
- package/dist/{index.js → index.cjs} +46 -15
- package/dist/{index.d.ts → index.d.cts} +24 -12
- package/dist/index.d.mts +191 -0
- package/dist/index.mjs +177 -0
- package/dist/keyCodes.d.mts +101 -0
- package/dist/keyCodes.mjs +201 -0
- package/node22.json +16 -0
- package/package.json +19 -9
- package/prebuilds/win32-x64/node-native-win-utils.node +0 -0
- package/src/cpp/keyboard.cpp +141 -140
- package/src/cpp/main.cpp +9 -3
- package/src/cpp/{capturewindow.cpp → screenshot.cpp} +174 -2
- package/src/dirnameLocal.mts +3 -0
- package/src/index.mts +454 -0
- package/src/keyCodes.mts +203 -0
- package/dllCopy.js +0 -14
- package/include/opencv2/core/affine.hpp +0 -678
- package/include/opencv2/core/async.hpp +0 -105
- package/include/opencv2/core/base.hpp +0 -664
- package/include/opencv2/core/bindings_utils.hpp +0 -325
- package/include/opencv2/core/bufferpool.hpp +0 -40
- package/include/opencv2/core/check.hpp +0 -170
- package/include/opencv2/core/core.hpp +0 -48
- package/include/opencv2/core/core_c.h +0 -3128
- package/include/opencv2/core/cuda/block.hpp +0 -211
- package/include/opencv2/core/cuda/border_interpolate.hpp +0 -722
- package/include/opencv2/core/cuda/color.hpp +0 -309
- package/include/opencv2/core/cuda/common.hpp +0 -131
- package/include/opencv2/core/cuda/datamov_utils.hpp +0 -113
- package/include/opencv2/core/cuda/detail/color_detail.hpp +0 -2018
- package/include/opencv2/core/cuda/detail/reduce.hpp +0 -365
- package/include/opencv2/core/cuda/detail/reduce_key_val.hpp +0 -502
- package/include/opencv2/core/cuda/detail/transform_detail.hpp +0 -392
- package/include/opencv2/core/cuda/detail/type_traits_detail.hpp +0 -191
- package/include/opencv2/core/cuda/detail/vec_distance_detail.hpp +0 -121
- package/include/opencv2/core/cuda/dynamic_smem.hpp +0 -88
- package/include/opencv2/core/cuda/emulation.hpp +0 -269
- package/include/opencv2/core/cuda/filters.hpp +0 -293
- package/include/opencv2/core/cuda/funcattrib.hpp +0 -79
- package/include/opencv2/core/cuda/functional.hpp +0 -805
- package/include/opencv2/core/cuda/limits.hpp +0 -128
- package/include/opencv2/core/cuda/reduce.hpp +0 -209
- package/include/opencv2/core/cuda/saturate_cast.hpp +0 -292
- package/include/opencv2/core/cuda/scan.hpp +0 -258
- package/include/opencv2/core/cuda/simd_functions.hpp +0 -869
- package/include/opencv2/core/cuda/transform.hpp +0 -75
- package/include/opencv2/core/cuda/type_traits.hpp +0 -90
- package/include/opencv2/core/cuda/utility.hpp +0 -230
- package/include/opencv2/core/cuda/vec_distance.hpp +0 -232
- package/include/opencv2/core/cuda/vec_math.hpp +0 -923
- package/include/opencv2/core/cuda/vec_traits.hpp +0 -288
- package/include/opencv2/core/cuda/warp.hpp +0 -139
- package/include/opencv2/core/cuda/warp_reduce.hpp +0 -76
- package/include/opencv2/core/cuda/warp_shuffle.hpp +0 -162
- package/include/opencv2/core/cuda.hpp +0 -1279
- package/include/opencv2/core/cuda.inl.hpp +0 -763
- package/include/opencv2/core/cuda_stream_accessor.hpp +0 -86
- package/include/opencv2/core/cuda_types.hpp +0 -144
- package/include/opencv2/core/cv_cpu_dispatch.h +0 -381
- package/include/opencv2/core/cv_cpu_helper.h +0 -550
- package/include/opencv2/core/cvdef.h +0 -973
- package/include/opencv2/core/cvstd.hpp +0 -190
- package/include/opencv2/core/cvstd.inl.hpp +0 -197
- package/include/opencv2/core/cvstd_wrapper.hpp +0 -154
- package/include/opencv2/core/detail/async_promise.hpp +0 -71
- package/include/opencv2/core/detail/dispatch_helper.impl.hpp +0 -49
- package/include/opencv2/core/detail/exception_ptr.hpp +0 -27
- package/include/opencv2/core/directx.hpp +0 -184
- package/include/opencv2/core/dualquaternion.hpp +0 -979
- package/include/opencv2/core/dualquaternion.inl.hpp +0 -487
- package/include/opencv2/core/eigen.hpp +0 -402
- package/include/opencv2/core/fast_math.hpp +0 -433
- package/include/opencv2/core/hal/hal.hpp +0 -256
- package/include/opencv2/core/hal/interface.h +0 -190
- package/include/opencv2/core/hal/intrin.hpp +0 -939
- package/include/opencv2/core/hal/intrin_avx.hpp +0 -3177
- package/include/opencv2/core/hal/intrin_avx512.hpp +0 -3090
- package/include/opencv2/core/hal/intrin_cpp.hpp +0 -3321
- package/include/opencv2/core/hal/intrin_forward.hpp +0 -191
- package/include/opencv2/core/hal/intrin_lasx.hpp +0 -3236
- package/include/opencv2/core/hal/intrin_msa.hpp +0 -1887
- package/include/opencv2/core/hal/intrin_neon.hpp +0 -2610
- package/include/opencv2/core/hal/intrin_rvv.hpp +0 -3320
- package/include/opencv2/core/hal/intrin_rvv071.hpp +0 -2545
- package/include/opencv2/core/hal/intrin_rvv_scalable.hpp +0 -2080
- package/include/opencv2/core/hal/intrin_sse.hpp +0 -3467
- package/include/opencv2/core/hal/intrin_sse_em.hpp +0 -180
- package/include/opencv2/core/hal/intrin_vsx.hpp +0 -1608
- package/include/opencv2/core/hal/intrin_wasm.hpp +0 -2782
- package/include/opencv2/core/hal/msa_macros.h +0 -1558
- package/include/opencv2/core/hal/simd_utils.impl.hpp +0 -186
- package/include/opencv2/core/llapi/llapi.h +0 -102
- package/include/opencv2/core/mat.hpp +0 -3775
- package/include/opencv2/core/mat.inl.hpp +0 -3422
- package/include/opencv2/core/matx.hpp +0 -1536
- package/include/opencv2/core/neon_utils.hpp +0 -128
- package/include/opencv2/core/ocl.hpp +0 -917
- package/include/opencv2/core/ocl_genbase.hpp +0 -69
- package/include/opencv2/core/opencl/ocl_defs.hpp +0 -82
- package/include/opencv2/core/opencl/opencl_info.hpp +0 -212
- package/include/opencv2/core/opencl/opencl_svm.hpp +0 -81
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_clblas.hpp +0 -602
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_clfft.hpp +0 -146
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_core.hpp +0 -371
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_core_wrappers.hpp +0 -272
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_gl.hpp +0 -62
- package/include/opencv2/core/opencl/runtime/autogenerated/opencl_gl_wrappers.hpp +0 -42
- package/include/opencv2/core/opencl/runtime/opencl_clblas.hpp +0 -53
- package/include/opencv2/core/opencl/runtime/opencl_clfft.hpp +0 -53
- package/include/opencv2/core/opencl/runtime/opencl_core.hpp +0 -84
- package/include/opencv2/core/opencl/runtime/opencl_core_wrappers.hpp +0 -47
- package/include/opencv2/core/opencl/runtime/opencl_gl.hpp +0 -53
- package/include/opencv2/core/opencl/runtime/opencl_gl_wrappers.hpp +0 -47
- package/include/opencv2/core/opencl/runtime/opencl_svm_20.hpp +0 -48
- package/include/opencv2/core/opencl/runtime/opencl_svm_definitions.hpp +0 -42
- package/include/opencv2/core/opencl/runtime/opencl_svm_hsa_extension.hpp +0 -166
- package/include/opencv2/core/opengl.hpp +0 -733
- package/include/opencv2/core/openvx/ovx_defs.hpp +0 -48
- package/include/opencv2/core/operations.hpp +0 -610
- package/include/opencv2/core/optim.hpp +0 -302
- package/include/opencv2/core/ovx.hpp +0 -28
- package/include/opencv2/core/parallel/backend/parallel_for.openmp.hpp +0 -72
- package/include/opencv2/core/parallel/backend/parallel_for.tbb.hpp +0 -153
- package/include/opencv2/core/parallel/parallel_backend.hpp +0 -90
- package/include/opencv2/core/persistence.hpp +0 -1350
- package/include/opencv2/core/private/cv_cpu_include_simd_declarations.hpp +0 -30
- package/include/opencv2/core/private.cuda.hpp +0 -169
- package/include/opencv2/core/private.hpp +0 -896
- package/include/opencv2/core/quaternion.hpp +0 -1696
- package/include/opencv2/core/quaternion.inl.hpp +0 -1063
- package/include/opencv2/core/saturate.hpp +0 -180
- package/include/opencv2/core/simd_intrinsics.hpp +0 -87
- package/include/opencv2/core/softfloat.hpp +0 -514
- package/include/opencv2/core/sse_utils.hpp +0 -652
- package/include/opencv2/core/traits.hpp +0 -417
- package/include/opencv2/core/types.hpp +0 -2457
- package/include/opencv2/core/types_c.h +0 -2126
- package/include/opencv2/core/utility.hpp +0 -1229
- package/include/opencv2/core/utils/allocator_stats.hpp +0 -29
- package/include/opencv2/core/utils/allocator_stats.impl.hpp +0 -158
- package/include/opencv2/core/utils/buffer_area.private.hpp +0 -136
- package/include/opencv2/core/utils/configuration.private.hpp +0 -22
- package/include/opencv2/core/utils/filesystem.hpp +0 -82
- package/include/opencv2/core/utils/filesystem.private.hpp +0 -66
- package/include/opencv2/core/utils/fp_control.private.hpp +0 -29
- package/include/opencv2/core/utils/fp_control_utils.hpp +0 -69
- package/include/opencv2/core/utils/instrumentation.hpp +0 -125
- package/include/opencv2/core/utils/lock.private.hpp +0 -119
- package/include/opencv2/core/utils/logger.defines.hpp +0 -42
- package/include/opencv2/core/utils/logger.hpp +0 -218
- package/include/opencv2/core/utils/logtag.hpp +0 -28
- package/include/opencv2/core/utils/plugin_loader.private.hpp +0 -165
- package/include/opencv2/core/utils/tls.hpp +0 -235
- package/include/opencv2/core/utils/trace.hpp +0 -252
- package/include/opencv2/core/utils/trace.private.hpp +0 -421
- package/include/opencv2/core/va_intel.hpp +0 -75
- package/include/opencv2/core/version.hpp +0 -26
- package/include/opencv2/core/vsx_utils.hpp +0 -1047
- package/include/opencv2/core.hpp +0 -3365
- package/include/opencv2/imgcodecs/imgcodecs.hpp +0 -48
- package/include/opencv2/imgcodecs/imgcodecs_c.h +0 -1
- package/include/opencv2/imgcodecs/ios.h +0 -59
- package/include/opencv2/imgcodecs/legacy/constants_c.h +0 -54
- package/include/opencv2/imgcodecs/macosx.h +0 -20
- package/include/opencv2/imgcodecs.hpp +0 -407
- package/include/opencv2/imgproc/bindings.hpp +0 -34
- package/include/opencv2/imgproc/detail/gcgraph.hpp +0 -395
- package/include/opencv2/imgproc/hal/hal.hpp +0 -246
- package/include/opencv2/imgproc/hal/interface.h +0 -46
- package/include/opencv2/imgproc/imgproc.hpp +0 -48
- package/include/opencv2/imgproc/imgproc_c.h +0 -1177
- package/include/opencv2/imgproc/segmentation.hpp +0 -141
- package/include/opencv2/imgproc/types_c.h +0 -659
- package/include/opencv2/imgproc.hpp +0 -5035
- package/include/opencv2/opencv_modules.hpp +0 -17
- package/include/tesseract/baseapi.h +0 -820
- package/include/tesseract/capi.h +0 -485
- package/include/tesseract/export.h +0 -37
- package/include/tesseract/ltrresultiterator.h +0 -235
- package/include/tesseract/ocrclass.h +0 -158
- package/include/tesseract/osdetect.h +0 -139
- package/include/tesseract/pageiterator.h +0 -364
- package/include/tesseract/publictypes.h +0 -281
- package/include/tesseract/renderer.h +0 -334
- package/include/tesseract/resultiterator.h +0 -250
- package/include/tesseract/tesseract.h +0 -40
- package/include/tesseract/unichar.h +0 -174
- package/include/tesseract/version.h +0 -34
- package/include/tesseract/version.h.in +0 -34
- package/libs/libjpeg-turbo.lib +0 -0
- package/libs/libpng.lib +0 -0
- package/libs/opencv_core470.lib +0 -0
- package/libs/opencv_imgcodecs470.lib +0 -0
- package/libs/opencv_imgproc470.lib +0 -0
- package/libs/tesseract.lib +0 -0
- package/libs/zlib.lib +0 -0
- /package/dist/{keyCodes.js → keyCodes.cjs} +0 -0
- /package/dist/{keyCodes.d.ts → keyCodes.d.cts} +0 -0
|
@@ -1,288 +0,0 @@
|
|
|
1
|
-
/*M///////////////////////////////////////////////////////////////////////////////////////
|
|
2
|
-
//
|
|
3
|
-
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
|
|
4
|
-
//
|
|
5
|
-
// By downloading, copying, installing or using the software you agree to this license.
|
|
6
|
-
// If you do not agree to this license, do not download, install,
|
|
7
|
-
// copy or use the software.
|
|
8
|
-
//
|
|
9
|
-
//
|
|
10
|
-
// License Agreement
|
|
11
|
-
// For Open Source Computer Vision Library
|
|
12
|
-
//
|
|
13
|
-
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
|
|
14
|
-
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
|
|
15
|
-
// Third party copyrights are property of their respective owners.
|
|
16
|
-
//
|
|
17
|
-
// Redistribution and use in source and binary forms, with or without modification,
|
|
18
|
-
// are permitted provided that the following conditions are met:
|
|
19
|
-
//
|
|
20
|
-
// * Redistribution's of source code must retain the above copyright notice,
|
|
21
|
-
// this list of conditions and the following disclaimer.
|
|
22
|
-
//
|
|
23
|
-
// * Redistribution's in binary form must reproduce the above copyright notice,
|
|
24
|
-
// this list of conditions and the following disclaimer in the documentation
|
|
25
|
-
// and/or other materials provided with the distribution.
|
|
26
|
-
//
|
|
27
|
-
// * The name of the copyright holders may not be used to endorse or promote products
|
|
28
|
-
// derived from this software without specific prior written permission.
|
|
29
|
-
//
|
|
30
|
-
// This software is provided by the copyright holders and contributors "as is" and
|
|
31
|
-
// any express or implied warranties, including, but not limited to, the implied
|
|
32
|
-
// warranties of merchantability and fitness for a particular purpose are disclaimed.
|
|
33
|
-
// In no event shall the Intel Corporation or contributors be liable for any direct,
|
|
34
|
-
// indirect, incidental, special, exemplary, or consequential damages
|
|
35
|
-
// (including, but not limited to, procurement of substitute goods or services;
|
|
36
|
-
// loss of use, data, or profits; or business interruption) however caused
|
|
37
|
-
// and on any theory of liability, whether in contract, strict liability,
|
|
38
|
-
// or tort (including negligence or otherwise) arising in any way out of
|
|
39
|
-
// the use of this software, even if advised of the possibility of such damage.
|
|
40
|
-
//
|
|
41
|
-
//M*/
|
|
42
|
-
|
|
43
|
-
#ifndef OPENCV_CUDA_VEC_TRAITS_HPP
|
|
44
|
-
#define OPENCV_CUDA_VEC_TRAITS_HPP
|
|
45
|
-
|
|
46
|
-
#include "common.hpp"
|
|
47
|
-
|
|
48
|
-
/** @file
|
|
49
|
-
* @deprecated Use @ref cudev instead.
|
|
50
|
-
*/
|
|
51
|
-
|
|
52
|
-
//! @cond IGNORED
|
|
53
|
-
|
|
54
|
-
namespace cv { namespace cuda { namespace device
|
|
55
|
-
{
|
|
56
|
-
template<typename T, int N> struct TypeVec;
|
|
57
|
-
|
|
58
|
-
struct __align__(8) uchar8
|
|
59
|
-
{
|
|
60
|
-
uchar a0, a1, a2, a3, a4, a5, a6, a7;
|
|
61
|
-
};
|
|
62
|
-
static __host__ __device__ __forceinline__ uchar8 make_uchar8(uchar a0, uchar a1, uchar a2, uchar a3, uchar a4, uchar a5, uchar a6, uchar a7)
|
|
63
|
-
{
|
|
64
|
-
uchar8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
65
|
-
return val;
|
|
66
|
-
}
|
|
67
|
-
struct __align__(8) char8
|
|
68
|
-
{
|
|
69
|
-
schar a0, a1, a2, a3, a4, a5, a6, a7;
|
|
70
|
-
};
|
|
71
|
-
static __host__ __device__ __forceinline__ char8 make_char8(schar a0, schar a1, schar a2, schar a3, schar a4, schar a5, schar a6, schar a7)
|
|
72
|
-
{
|
|
73
|
-
char8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
74
|
-
return val;
|
|
75
|
-
}
|
|
76
|
-
struct __align__(16) ushort8
|
|
77
|
-
{
|
|
78
|
-
ushort a0, a1, a2, a3, a4, a5, a6, a7;
|
|
79
|
-
};
|
|
80
|
-
static __host__ __device__ __forceinline__ ushort8 make_ushort8(ushort a0, ushort a1, ushort a2, ushort a3, ushort a4, ushort a5, ushort a6, ushort a7)
|
|
81
|
-
{
|
|
82
|
-
ushort8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
83
|
-
return val;
|
|
84
|
-
}
|
|
85
|
-
struct __align__(16) short8
|
|
86
|
-
{
|
|
87
|
-
short a0, a1, a2, a3, a4, a5, a6, a7;
|
|
88
|
-
};
|
|
89
|
-
static __host__ __device__ __forceinline__ short8 make_short8(short a0, short a1, short a2, short a3, short a4, short a5, short a6, short a7)
|
|
90
|
-
{
|
|
91
|
-
short8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
92
|
-
return val;
|
|
93
|
-
}
|
|
94
|
-
struct __align__(32) uint8
|
|
95
|
-
{
|
|
96
|
-
uint a0, a1, a2, a3, a4, a5, a6, a7;
|
|
97
|
-
};
|
|
98
|
-
static __host__ __device__ __forceinline__ uint8 make_uint8(uint a0, uint a1, uint a2, uint a3, uint a4, uint a5, uint a6, uint a7)
|
|
99
|
-
{
|
|
100
|
-
uint8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
101
|
-
return val;
|
|
102
|
-
}
|
|
103
|
-
struct __align__(32) int8
|
|
104
|
-
{
|
|
105
|
-
int a0, a1, a2, a3, a4, a5, a6, a7;
|
|
106
|
-
};
|
|
107
|
-
static __host__ __device__ __forceinline__ int8 make_int8(int a0, int a1, int a2, int a3, int a4, int a5, int a6, int a7)
|
|
108
|
-
{
|
|
109
|
-
int8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
110
|
-
return val;
|
|
111
|
-
}
|
|
112
|
-
struct __align__(32) float8
|
|
113
|
-
{
|
|
114
|
-
float a0, a1, a2, a3, a4, a5, a6, a7;
|
|
115
|
-
};
|
|
116
|
-
static __host__ __device__ __forceinline__ float8 make_float8(float a0, float a1, float a2, float a3, float a4, float a5, float a6, float a7)
|
|
117
|
-
{
|
|
118
|
-
float8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
119
|
-
return val;
|
|
120
|
-
}
|
|
121
|
-
struct double8
|
|
122
|
-
{
|
|
123
|
-
double a0, a1, a2, a3, a4, a5, a6, a7;
|
|
124
|
-
};
|
|
125
|
-
static __host__ __device__ __forceinline__ double8 make_double8(double a0, double a1, double a2, double a3, double a4, double a5, double a6, double a7)
|
|
126
|
-
{
|
|
127
|
-
double8 val = {a0, a1, a2, a3, a4, a5, a6, a7};
|
|
128
|
-
return val;
|
|
129
|
-
}
|
|
130
|
-
|
|
131
|
-
#define OPENCV_CUDA_IMPLEMENT_TYPE_VEC(type) \
|
|
132
|
-
template<> struct TypeVec<type, 1> { typedef type vec_type; }; \
|
|
133
|
-
template<> struct TypeVec<type ## 1, 1> { typedef type ## 1 vec_type; }; \
|
|
134
|
-
template<> struct TypeVec<type, 2> { typedef type ## 2 vec_type; }; \
|
|
135
|
-
template<> struct TypeVec<type ## 2, 2> { typedef type ## 2 vec_type; }; \
|
|
136
|
-
template<> struct TypeVec<type, 3> { typedef type ## 3 vec_type; }; \
|
|
137
|
-
template<> struct TypeVec<type ## 3, 3> { typedef type ## 3 vec_type; }; \
|
|
138
|
-
template<> struct TypeVec<type, 4> { typedef type ## 4 vec_type; }; \
|
|
139
|
-
template<> struct TypeVec<type ## 4, 4> { typedef type ## 4 vec_type; }; \
|
|
140
|
-
template<> struct TypeVec<type, 8> { typedef type ## 8 vec_type; }; \
|
|
141
|
-
template<> struct TypeVec<type ## 8, 8> { typedef type ## 8 vec_type; };
|
|
142
|
-
|
|
143
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(uchar)
|
|
144
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(char)
|
|
145
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(ushort)
|
|
146
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(short)
|
|
147
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(int)
|
|
148
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(uint)
|
|
149
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(float)
|
|
150
|
-
OPENCV_CUDA_IMPLEMENT_TYPE_VEC(double)
|
|
151
|
-
|
|
152
|
-
#undef OPENCV_CUDA_IMPLEMENT_TYPE_VEC
|
|
153
|
-
|
|
154
|
-
template<> struct TypeVec<schar, 1> { typedef schar vec_type; };
|
|
155
|
-
template<> struct TypeVec<schar, 2> { typedef char2 vec_type; };
|
|
156
|
-
template<> struct TypeVec<schar, 3> { typedef char3 vec_type; };
|
|
157
|
-
template<> struct TypeVec<schar, 4> { typedef char4 vec_type; };
|
|
158
|
-
template<> struct TypeVec<schar, 8> { typedef char8 vec_type; };
|
|
159
|
-
|
|
160
|
-
template<> struct TypeVec<bool, 1> { typedef uchar vec_type; };
|
|
161
|
-
template<> struct TypeVec<bool, 2> { typedef uchar2 vec_type; };
|
|
162
|
-
template<> struct TypeVec<bool, 3> { typedef uchar3 vec_type; };
|
|
163
|
-
template<> struct TypeVec<bool, 4> { typedef uchar4 vec_type; };
|
|
164
|
-
template<> struct TypeVec<bool, 8> { typedef uchar8 vec_type; };
|
|
165
|
-
|
|
166
|
-
template<typename T> struct VecTraits;
|
|
167
|
-
|
|
168
|
-
#define OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(type) \
|
|
169
|
-
template<> struct VecTraits<type> \
|
|
170
|
-
{ \
|
|
171
|
-
typedef type elem_type; \
|
|
172
|
-
enum {cn=1}; \
|
|
173
|
-
static __device__ __host__ __forceinline__ type all(type v) {return v;} \
|
|
174
|
-
static __device__ __host__ __forceinline__ type make(type x) {return x;} \
|
|
175
|
-
static __device__ __host__ __forceinline__ type make(const type* v) {return *v;} \
|
|
176
|
-
}; \
|
|
177
|
-
template<> struct VecTraits<type ## 1> \
|
|
178
|
-
{ \
|
|
179
|
-
typedef type elem_type; \
|
|
180
|
-
enum {cn=1}; \
|
|
181
|
-
static __device__ __host__ __forceinline__ type ## 1 all(type v) {return make_ ## type ## 1(v);} \
|
|
182
|
-
static __device__ __host__ __forceinline__ type ## 1 make(type x) {return make_ ## type ## 1(x);} \
|
|
183
|
-
static __device__ __host__ __forceinline__ type ## 1 make(const type* v) {return make_ ## type ## 1(*v);} \
|
|
184
|
-
}; \
|
|
185
|
-
template<> struct VecTraits<type ## 2> \
|
|
186
|
-
{ \
|
|
187
|
-
typedef type elem_type; \
|
|
188
|
-
enum {cn=2}; \
|
|
189
|
-
static __device__ __host__ __forceinline__ type ## 2 all(type v) {return make_ ## type ## 2(v, v);} \
|
|
190
|
-
static __device__ __host__ __forceinline__ type ## 2 make(type x, type y) {return make_ ## type ## 2(x, y);} \
|
|
191
|
-
static __device__ __host__ __forceinline__ type ## 2 make(const type* v) {return make_ ## type ## 2(v[0], v[1]);} \
|
|
192
|
-
}; \
|
|
193
|
-
template<> struct VecTraits<type ## 3> \
|
|
194
|
-
{ \
|
|
195
|
-
typedef type elem_type; \
|
|
196
|
-
enum {cn=3}; \
|
|
197
|
-
static __device__ __host__ __forceinline__ type ## 3 all(type v) {return make_ ## type ## 3(v, v, v);} \
|
|
198
|
-
static __device__ __host__ __forceinline__ type ## 3 make(type x, type y, type z) {return make_ ## type ## 3(x, y, z);} \
|
|
199
|
-
static __device__ __host__ __forceinline__ type ## 3 make(const type* v) {return make_ ## type ## 3(v[0], v[1], v[2]);} \
|
|
200
|
-
}; \
|
|
201
|
-
template<> struct VecTraits<type ## 4> \
|
|
202
|
-
{ \
|
|
203
|
-
typedef type elem_type; \
|
|
204
|
-
enum {cn=4}; \
|
|
205
|
-
static __device__ __host__ __forceinline__ type ## 4 all(type v) {return make_ ## type ## 4(v, v, v, v);} \
|
|
206
|
-
static __device__ __host__ __forceinline__ type ## 4 make(type x, type y, type z, type w) {return make_ ## type ## 4(x, y, z, w);} \
|
|
207
|
-
static __device__ __host__ __forceinline__ type ## 4 make(const type* v) {return make_ ## type ## 4(v[0], v[1], v[2], v[3]);} \
|
|
208
|
-
}; \
|
|
209
|
-
template<> struct VecTraits<type ## 8> \
|
|
210
|
-
{ \
|
|
211
|
-
typedef type elem_type; \
|
|
212
|
-
enum {cn=8}; \
|
|
213
|
-
static __device__ __host__ __forceinline__ type ## 8 all(type v) {return make_ ## type ## 8(v, v, v, v, v, v, v, v);} \
|
|
214
|
-
static __device__ __host__ __forceinline__ type ## 8 make(type a0, type a1, type a2, type a3, type a4, type a5, type a6, type a7) {return make_ ## type ## 8(a0, a1, a2, a3, a4, a5, a6, a7);} \
|
|
215
|
-
static __device__ __host__ __forceinline__ type ## 8 make(const type* v) {return make_ ## type ## 8(v[0], v[1], v[2], v[3], v[4], v[5], v[6], v[7]);} \
|
|
216
|
-
};
|
|
217
|
-
|
|
218
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(uchar)
|
|
219
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(ushort)
|
|
220
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(short)
|
|
221
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(int)
|
|
222
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(uint)
|
|
223
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(float)
|
|
224
|
-
OPENCV_CUDA_IMPLEMENT_VEC_TRAITS(double)
|
|
225
|
-
|
|
226
|
-
#undef OPENCV_CUDA_IMPLEMENT_VEC_TRAITS
|
|
227
|
-
|
|
228
|
-
template<> struct VecTraits<char>
|
|
229
|
-
{
|
|
230
|
-
typedef char elem_type;
|
|
231
|
-
enum {cn=1};
|
|
232
|
-
static __device__ __host__ __forceinline__ char all(char v) {return v;}
|
|
233
|
-
static __device__ __host__ __forceinline__ char make(char x) {return x;}
|
|
234
|
-
static __device__ __host__ __forceinline__ char make(const char* x) {return *x;}
|
|
235
|
-
};
|
|
236
|
-
template<> struct VecTraits<schar>
|
|
237
|
-
{
|
|
238
|
-
typedef schar elem_type;
|
|
239
|
-
enum {cn=1};
|
|
240
|
-
static __device__ __host__ __forceinline__ schar all(schar v) {return v;}
|
|
241
|
-
static __device__ __host__ __forceinline__ schar make(schar x) {return x;}
|
|
242
|
-
static __device__ __host__ __forceinline__ schar make(const schar* x) {return *x;}
|
|
243
|
-
};
|
|
244
|
-
template<> struct VecTraits<char1>
|
|
245
|
-
{
|
|
246
|
-
typedef schar elem_type;
|
|
247
|
-
enum {cn=1};
|
|
248
|
-
static __device__ __host__ __forceinline__ char1 all(schar v) {return make_char1(v);}
|
|
249
|
-
static __device__ __host__ __forceinline__ char1 make(schar x) {return make_char1(x);}
|
|
250
|
-
static __device__ __host__ __forceinline__ char1 make(const schar* v) {return make_char1(v[0]);}
|
|
251
|
-
};
|
|
252
|
-
template<> struct VecTraits<char2>
|
|
253
|
-
{
|
|
254
|
-
typedef schar elem_type;
|
|
255
|
-
enum {cn=2};
|
|
256
|
-
static __device__ __host__ __forceinline__ char2 all(schar v) {return make_char2(v, v);}
|
|
257
|
-
static __device__ __host__ __forceinline__ char2 make(schar x, schar y) {return make_char2(x, y);}
|
|
258
|
-
static __device__ __host__ __forceinline__ char2 make(const schar* v) {return make_char2(v[0], v[1]);}
|
|
259
|
-
};
|
|
260
|
-
template<> struct VecTraits<char3>
|
|
261
|
-
{
|
|
262
|
-
typedef schar elem_type;
|
|
263
|
-
enum {cn=3};
|
|
264
|
-
static __device__ __host__ __forceinline__ char3 all(schar v) {return make_char3(v, v, v);}
|
|
265
|
-
static __device__ __host__ __forceinline__ char3 make(schar x, schar y, schar z) {return make_char3(x, y, z);}
|
|
266
|
-
static __device__ __host__ __forceinline__ char3 make(const schar* v) {return make_char3(v[0], v[1], v[2]);}
|
|
267
|
-
};
|
|
268
|
-
template<> struct VecTraits<char4>
|
|
269
|
-
{
|
|
270
|
-
typedef schar elem_type;
|
|
271
|
-
enum {cn=4};
|
|
272
|
-
static __device__ __host__ __forceinline__ char4 all(schar v) {return make_char4(v, v, v, v);}
|
|
273
|
-
static __device__ __host__ __forceinline__ char4 make(schar x, schar y, schar z, schar w) {return make_char4(x, y, z, w);}
|
|
274
|
-
static __device__ __host__ __forceinline__ char4 make(const schar* v) {return make_char4(v[0], v[1], v[2], v[3]);}
|
|
275
|
-
};
|
|
276
|
-
template<> struct VecTraits<char8>
|
|
277
|
-
{
|
|
278
|
-
typedef schar elem_type;
|
|
279
|
-
enum {cn=8};
|
|
280
|
-
static __device__ __host__ __forceinline__ char8 all(schar v) {return make_char8(v, v, v, v, v, v, v, v);}
|
|
281
|
-
static __device__ __host__ __forceinline__ char8 make(schar a0, schar a1, schar a2, schar a3, schar a4, schar a5, schar a6, schar a7) {return make_char8(a0, a1, a2, a3, a4, a5, a6, a7);}
|
|
282
|
-
static __device__ __host__ __forceinline__ char8 make(const schar* v) {return make_char8(v[0], v[1], v[2], v[3], v[4], v[5], v[6], v[7]);}
|
|
283
|
-
};
|
|
284
|
-
}}} // namespace cv { namespace cuda { namespace cudev
|
|
285
|
-
|
|
286
|
-
//! @endcond
|
|
287
|
-
|
|
288
|
-
#endif // OPENCV_CUDA_VEC_TRAITS_HPP
|
|
@@ -1,139 +0,0 @@
|
|
|
1
|
-
/*M///////////////////////////////////////////////////////////////////////////////////////
|
|
2
|
-
//
|
|
3
|
-
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
|
|
4
|
-
//
|
|
5
|
-
// By downloading, copying, installing or using the software you agree to this license.
|
|
6
|
-
// If you do not agree to this license, do not download, install,
|
|
7
|
-
// copy or use the software.
|
|
8
|
-
//
|
|
9
|
-
//
|
|
10
|
-
// License Agreement
|
|
11
|
-
// For Open Source Computer Vision Library
|
|
12
|
-
//
|
|
13
|
-
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
|
|
14
|
-
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
|
|
15
|
-
// Third party copyrights are property of their respective owners.
|
|
16
|
-
//
|
|
17
|
-
// Redistribution and use in source and binary forms, with or without modification,
|
|
18
|
-
// are permitted provided that the following conditions are met:
|
|
19
|
-
//
|
|
20
|
-
// * Redistribution's of source code must retain the above copyright notice,
|
|
21
|
-
// this list of conditions and the following disclaimer.
|
|
22
|
-
//
|
|
23
|
-
// * Redistribution's in binary form must reproduce the above copyright notice,
|
|
24
|
-
// this list of conditions and the following disclaimer in the documentation
|
|
25
|
-
// and/or other materials provided with the distribution.
|
|
26
|
-
//
|
|
27
|
-
// * The name of the copyright holders may not be used to endorse or promote products
|
|
28
|
-
// derived from this software without specific prior written permission.
|
|
29
|
-
//
|
|
30
|
-
// This software is provided by the copyright holders and contributors "as is" and
|
|
31
|
-
// any express or implied warranties, including, but not limited to, the implied
|
|
32
|
-
// warranties of merchantability and fitness for a particular purpose are disclaimed.
|
|
33
|
-
// In no event shall the Intel Corporation or contributors be liable for any direct,
|
|
34
|
-
// indirect, incidental, special, exemplary, or consequential damages
|
|
35
|
-
// (including, but not limited to, procurement of substitute goods or services;
|
|
36
|
-
// loss of use, data, or profits; or business interruption) however caused
|
|
37
|
-
// and on any theory of liability, whether in contract, strict liability,
|
|
38
|
-
// or tort (including negligence or otherwise) arising in any way out of
|
|
39
|
-
// the use of this software, even if advised of the possibility of such damage.
|
|
40
|
-
//
|
|
41
|
-
//M*/
|
|
42
|
-
|
|
43
|
-
#ifndef OPENCV_CUDA_DEVICE_WARP_HPP
|
|
44
|
-
#define OPENCV_CUDA_DEVICE_WARP_HPP
|
|
45
|
-
|
|
46
|
-
/** @file
|
|
47
|
-
* @deprecated Use @ref cudev instead.
|
|
48
|
-
*/
|
|
49
|
-
|
|
50
|
-
//! @cond IGNORED
|
|
51
|
-
|
|
52
|
-
namespace cv { namespace cuda { namespace device
|
|
53
|
-
{
|
|
54
|
-
struct Warp
|
|
55
|
-
{
|
|
56
|
-
enum
|
|
57
|
-
{
|
|
58
|
-
LOG_WARP_SIZE = 5,
|
|
59
|
-
WARP_SIZE = 1 << LOG_WARP_SIZE,
|
|
60
|
-
STRIDE = WARP_SIZE
|
|
61
|
-
};
|
|
62
|
-
|
|
63
|
-
/** \brief Returns the warp lane ID of the calling thread. */
|
|
64
|
-
static __device__ __forceinline__ unsigned int laneId()
|
|
65
|
-
{
|
|
66
|
-
unsigned int ret;
|
|
67
|
-
asm("mov.u32 %0, %%laneid;" : "=r"(ret) );
|
|
68
|
-
return ret;
|
|
69
|
-
}
|
|
70
|
-
|
|
71
|
-
template<typename It, typename T>
|
|
72
|
-
static __device__ __forceinline__ void fill(It beg, It end, const T& value)
|
|
73
|
-
{
|
|
74
|
-
for(It t = beg + laneId(); t < end; t += STRIDE)
|
|
75
|
-
*t = value;
|
|
76
|
-
}
|
|
77
|
-
|
|
78
|
-
template<typename InIt, typename OutIt>
|
|
79
|
-
static __device__ __forceinline__ OutIt copy(InIt beg, InIt end, OutIt out)
|
|
80
|
-
{
|
|
81
|
-
for(InIt t = beg + laneId(); t < end; t += STRIDE, out += STRIDE)
|
|
82
|
-
*out = *t;
|
|
83
|
-
return out;
|
|
84
|
-
}
|
|
85
|
-
|
|
86
|
-
template<typename InIt, typename OutIt, class UnOp>
|
|
87
|
-
static __device__ __forceinline__ OutIt transform(InIt beg, InIt end, OutIt out, UnOp op)
|
|
88
|
-
{
|
|
89
|
-
for(InIt t = beg + laneId(); t < end; t += STRIDE, out += STRIDE)
|
|
90
|
-
*out = op(*t);
|
|
91
|
-
return out;
|
|
92
|
-
}
|
|
93
|
-
|
|
94
|
-
template<typename InIt1, typename InIt2, typename OutIt, class BinOp>
|
|
95
|
-
static __device__ __forceinline__ OutIt transform(InIt1 beg1, InIt1 end1, InIt2 beg2, OutIt out, BinOp op)
|
|
96
|
-
{
|
|
97
|
-
unsigned int lane = laneId();
|
|
98
|
-
|
|
99
|
-
InIt1 t1 = beg1 + lane;
|
|
100
|
-
InIt2 t2 = beg2 + lane;
|
|
101
|
-
for(; t1 < end1; t1 += STRIDE, t2 += STRIDE, out += STRIDE)
|
|
102
|
-
*out = op(*t1, *t2);
|
|
103
|
-
return out;
|
|
104
|
-
}
|
|
105
|
-
|
|
106
|
-
template <class T, class BinOp>
|
|
107
|
-
static __device__ __forceinline__ T reduce(volatile T *ptr, BinOp op)
|
|
108
|
-
{
|
|
109
|
-
const unsigned int lane = laneId();
|
|
110
|
-
|
|
111
|
-
if (lane < 16)
|
|
112
|
-
{
|
|
113
|
-
T partial = ptr[lane];
|
|
114
|
-
|
|
115
|
-
ptr[lane] = partial = op(partial, ptr[lane + 16]);
|
|
116
|
-
ptr[lane] = partial = op(partial, ptr[lane + 8]);
|
|
117
|
-
ptr[lane] = partial = op(partial, ptr[lane + 4]);
|
|
118
|
-
ptr[lane] = partial = op(partial, ptr[lane + 2]);
|
|
119
|
-
ptr[lane] = partial = op(partial, ptr[lane + 1]);
|
|
120
|
-
}
|
|
121
|
-
|
|
122
|
-
return *ptr;
|
|
123
|
-
}
|
|
124
|
-
|
|
125
|
-
template<typename OutIt, typename T>
|
|
126
|
-
static __device__ __forceinline__ void yota(OutIt beg, OutIt end, T value)
|
|
127
|
-
{
|
|
128
|
-
unsigned int lane = laneId();
|
|
129
|
-
value += lane;
|
|
130
|
-
|
|
131
|
-
for(OutIt t = beg + lane; t < end; t += STRIDE, value += STRIDE)
|
|
132
|
-
*t = value;
|
|
133
|
-
}
|
|
134
|
-
};
|
|
135
|
-
}}} // namespace cv { namespace cuda { namespace cudev
|
|
136
|
-
|
|
137
|
-
//! @endcond
|
|
138
|
-
|
|
139
|
-
#endif /* OPENCV_CUDA_DEVICE_WARP_HPP */
|
|
@@ -1,76 +0,0 @@
|
|
|
1
|
-
/*M///////////////////////////////////////////////////////////////////////////////////////
|
|
2
|
-
//
|
|
3
|
-
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
|
|
4
|
-
//
|
|
5
|
-
// By downloading, copying, installing or using the software you agree to this license.
|
|
6
|
-
// If you do not agree to this license, do not download, install,
|
|
7
|
-
// copy or use the software.
|
|
8
|
-
//
|
|
9
|
-
//
|
|
10
|
-
// License Agreement
|
|
11
|
-
// For Open Source Computer Vision Library
|
|
12
|
-
//
|
|
13
|
-
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
|
|
14
|
-
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
|
|
15
|
-
// Third party copyrights are property of their respective owners.
|
|
16
|
-
//
|
|
17
|
-
// Redistribution and use in source and binary forms, with or without modification,
|
|
18
|
-
// are permitted provided that the following conditions are met:
|
|
19
|
-
//
|
|
20
|
-
// * Redistribution's of source code must retain the above copyright notice,
|
|
21
|
-
// this list of conditions and the following disclaimer.
|
|
22
|
-
//
|
|
23
|
-
// * Redistribution's in binary form must reproduce the above copyright notice,
|
|
24
|
-
// this list of conditions and the following disclaimer in the documentation
|
|
25
|
-
// and/or other materials provided with the distribution.
|
|
26
|
-
//
|
|
27
|
-
// * The name of the copyright holders may not be used to endorse or promote products
|
|
28
|
-
// derived from this software without specific prior written permission.
|
|
29
|
-
//
|
|
30
|
-
// This software is provided by the copyright holders and contributors "as is" and
|
|
31
|
-
// any express or implied warranties, including, but not limited to, the implied
|
|
32
|
-
// warranties of merchantability and fitness for a particular purpose are disclaimed.
|
|
33
|
-
// In no event shall the Intel Corporation or contributors be liable for any direct,
|
|
34
|
-
// indirect, incidental, special, exemplary, or consequential damages
|
|
35
|
-
// (including, but not limited to, procurement of substitute goods or services;
|
|
36
|
-
// loss of use, data, or profits; or business interruption) however caused
|
|
37
|
-
// and on any theory of liability, whether in contract, strict liability,
|
|
38
|
-
// or tort (including negligence or otherwise) arising in any way out of
|
|
39
|
-
// the use of this software, even if advised of the possibility of such damage.
|
|
40
|
-
//
|
|
41
|
-
//M*/
|
|
42
|
-
|
|
43
|
-
#ifndef OPENCV_CUDA_WARP_REDUCE_HPP__
|
|
44
|
-
#define OPENCV_CUDA_WARP_REDUCE_HPP__
|
|
45
|
-
|
|
46
|
-
/** @file
|
|
47
|
-
* @deprecated Use @ref cudev instead.
|
|
48
|
-
*/
|
|
49
|
-
|
|
50
|
-
//! @cond IGNORED
|
|
51
|
-
|
|
52
|
-
namespace cv { namespace cuda { namespace device
|
|
53
|
-
{
|
|
54
|
-
template <class T>
|
|
55
|
-
__device__ __forceinline__ T warp_reduce(volatile T *ptr , const unsigned int tid = threadIdx.x)
|
|
56
|
-
{
|
|
57
|
-
const unsigned int lane = tid & 31; // index of thread in warp (0..31)
|
|
58
|
-
|
|
59
|
-
if (lane < 16)
|
|
60
|
-
{
|
|
61
|
-
T partial = ptr[tid];
|
|
62
|
-
|
|
63
|
-
ptr[tid] = partial = partial + ptr[tid + 16];
|
|
64
|
-
ptr[tid] = partial = partial + ptr[tid + 8];
|
|
65
|
-
ptr[tid] = partial = partial + ptr[tid + 4];
|
|
66
|
-
ptr[tid] = partial = partial + ptr[tid + 2];
|
|
67
|
-
ptr[tid] = partial = partial + ptr[tid + 1];
|
|
68
|
-
}
|
|
69
|
-
|
|
70
|
-
return ptr[tid - lane];
|
|
71
|
-
}
|
|
72
|
-
}}} // namespace cv { namespace cuda { namespace cudev {
|
|
73
|
-
|
|
74
|
-
//! @endcond
|
|
75
|
-
|
|
76
|
-
#endif /* OPENCV_CUDA_WARP_REDUCE_HPP__ */
|
|
@@ -1,162 +0,0 @@
|
|
|
1
|
-
/*M///////////////////////////////////////////////////////////////////////////////////////
|
|
2
|
-
//
|
|
3
|
-
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
|
|
4
|
-
//
|
|
5
|
-
// By downloading, copying, installing or using the software you agree to this license.
|
|
6
|
-
// If you do not agree to this license, do not download, install,
|
|
7
|
-
// copy or use the software.
|
|
8
|
-
//
|
|
9
|
-
//
|
|
10
|
-
// License Agreement
|
|
11
|
-
// For Open Source Computer Vision Library
|
|
12
|
-
//
|
|
13
|
-
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
|
|
14
|
-
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
|
|
15
|
-
// Third party copyrights are property of their respective owners.
|
|
16
|
-
//
|
|
17
|
-
// Redistribution and use in source and binary forms, with or without modification,
|
|
18
|
-
// are permitted provided that the following conditions are met:
|
|
19
|
-
//
|
|
20
|
-
// * Redistribution's of source code must retain the above copyright notice,
|
|
21
|
-
// this list of conditions and the following disclaimer.
|
|
22
|
-
//
|
|
23
|
-
// * Redistribution's in binary form must reproduce the above copyright notice,
|
|
24
|
-
// this list of conditions and the following disclaimer in the documentation
|
|
25
|
-
// and/or other materials provided with the distribution.
|
|
26
|
-
//
|
|
27
|
-
// * The name of the copyright holders may not be used to endorse or promote products
|
|
28
|
-
// derived from this software without specific prior written permission.
|
|
29
|
-
//
|
|
30
|
-
// This software is provided by the copyright holders and contributors "as is" and
|
|
31
|
-
// any express or implied warranties, including, but not limited to, the implied
|
|
32
|
-
// warranties of merchantability and fitness for a particular purpose are disclaimed.
|
|
33
|
-
// In no event shall the Intel Corporation or contributors be liable for any direct,
|
|
34
|
-
// indirect, incidental, special, exemplary, or consequential damages
|
|
35
|
-
// (including, but not limited to, procurement of substitute goods or services;
|
|
36
|
-
// loss of use, data, or profits; or business interruption) however caused
|
|
37
|
-
// and on any theory of liability, whether in contract, strict liability,
|
|
38
|
-
// or tort (including negligence or otherwise) arising in any way out of
|
|
39
|
-
// the use of this software, even if advised of the possibility of such damage.
|
|
40
|
-
//
|
|
41
|
-
//M*/
|
|
42
|
-
|
|
43
|
-
#ifndef OPENCV_CUDA_WARP_SHUFFLE_HPP
|
|
44
|
-
#define OPENCV_CUDA_WARP_SHUFFLE_HPP
|
|
45
|
-
|
|
46
|
-
/** @file
|
|
47
|
-
* @deprecated Use @ref cudev instead.
|
|
48
|
-
*/
|
|
49
|
-
|
|
50
|
-
//! @cond IGNORED
|
|
51
|
-
|
|
52
|
-
namespace cv { namespace cuda { namespace device
|
|
53
|
-
{
|
|
54
|
-
#if __CUDACC_VER_MAJOR__ >= 9
|
|
55
|
-
# define __shfl(x, y, z) __shfl_sync(0xFFFFFFFFU, x, y, z)
|
|
56
|
-
# define __shfl_up(x, y, z) __shfl_up_sync(0xFFFFFFFFU, x, y, z)
|
|
57
|
-
# define __shfl_down(x, y, z) __shfl_down_sync(0xFFFFFFFFU, x, y, z)
|
|
58
|
-
#endif
|
|
59
|
-
template <typename T>
|
|
60
|
-
__device__ __forceinline__ T shfl(T val, int srcLane, int width = warpSize)
|
|
61
|
-
{
|
|
62
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
63
|
-
return __shfl(val, srcLane, width);
|
|
64
|
-
#else
|
|
65
|
-
return T();
|
|
66
|
-
#endif
|
|
67
|
-
}
|
|
68
|
-
__device__ __forceinline__ unsigned int shfl(unsigned int val, int srcLane, int width = warpSize)
|
|
69
|
-
{
|
|
70
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
71
|
-
return (unsigned int) __shfl((int) val, srcLane, width);
|
|
72
|
-
#else
|
|
73
|
-
return 0;
|
|
74
|
-
#endif
|
|
75
|
-
}
|
|
76
|
-
__device__ __forceinline__ double shfl(double val, int srcLane, int width = warpSize)
|
|
77
|
-
{
|
|
78
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
79
|
-
int lo = __double2loint(val);
|
|
80
|
-
int hi = __double2hiint(val);
|
|
81
|
-
|
|
82
|
-
lo = __shfl(lo, srcLane, width);
|
|
83
|
-
hi = __shfl(hi, srcLane, width);
|
|
84
|
-
|
|
85
|
-
return __hiloint2double(hi, lo);
|
|
86
|
-
#else
|
|
87
|
-
return 0.0;
|
|
88
|
-
#endif
|
|
89
|
-
}
|
|
90
|
-
|
|
91
|
-
template <typename T>
|
|
92
|
-
__device__ __forceinline__ T shfl_down(T val, unsigned int delta, int width = warpSize)
|
|
93
|
-
{
|
|
94
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
95
|
-
return __shfl_down(val, delta, width);
|
|
96
|
-
#else
|
|
97
|
-
return T();
|
|
98
|
-
#endif
|
|
99
|
-
}
|
|
100
|
-
__device__ __forceinline__ unsigned int shfl_down(unsigned int val, unsigned int delta, int width = warpSize)
|
|
101
|
-
{
|
|
102
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
103
|
-
return (unsigned int) __shfl_down((int) val, delta, width);
|
|
104
|
-
#else
|
|
105
|
-
return 0;
|
|
106
|
-
#endif
|
|
107
|
-
}
|
|
108
|
-
__device__ __forceinline__ double shfl_down(double val, unsigned int delta, int width = warpSize)
|
|
109
|
-
{
|
|
110
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
111
|
-
int lo = __double2loint(val);
|
|
112
|
-
int hi = __double2hiint(val);
|
|
113
|
-
|
|
114
|
-
lo = __shfl_down(lo, delta, width);
|
|
115
|
-
hi = __shfl_down(hi, delta, width);
|
|
116
|
-
|
|
117
|
-
return __hiloint2double(hi, lo);
|
|
118
|
-
#else
|
|
119
|
-
return 0.0;
|
|
120
|
-
#endif
|
|
121
|
-
}
|
|
122
|
-
|
|
123
|
-
template <typename T>
|
|
124
|
-
__device__ __forceinline__ T shfl_up(T val, unsigned int delta, int width = warpSize)
|
|
125
|
-
{
|
|
126
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
127
|
-
return __shfl_up(val, delta, width);
|
|
128
|
-
#else
|
|
129
|
-
return T();
|
|
130
|
-
#endif
|
|
131
|
-
}
|
|
132
|
-
__device__ __forceinline__ unsigned int shfl_up(unsigned int val, unsigned int delta, int width = warpSize)
|
|
133
|
-
{
|
|
134
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
135
|
-
return (unsigned int) __shfl_up((int) val, delta, width);
|
|
136
|
-
#else
|
|
137
|
-
return 0;
|
|
138
|
-
#endif
|
|
139
|
-
}
|
|
140
|
-
__device__ __forceinline__ double shfl_up(double val, unsigned int delta, int width = warpSize)
|
|
141
|
-
{
|
|
142
|
-
#if defined __CUDA_ARCH__ && __CUDA_ARCH__ >= 300
|
|
143
|
-
int lo = __double2loint(val);
|
|
144
|
-
int hi = __double2hiint(val);
|
|
145
|
-
|
|
146
|
-
lo = __shfl_up(lo, delta, width);
|
|
147
|
-
hi = __shfl_up(hi, delta, width);
|
|
148
|
-
|
|
149
|
-
return __hiloint2double(hi, lo);
|
|
150
|
-
#else
|
|
151
|
-
return 0.0;
|
|
152
|
-
#endif
|
|
153
|
-
}
|
|
154
|
-
}}}
|
|
155
|
-
|
|
156
|
-
# undef __shfl
|
|
157
|
-
# undef __shfl_up
|
|
158
|
-
# undef __shfl_down
|
|
159
|
-
|
|
160
|
-
//! @endcond
|
|
161
|
-
|
|
162
|
-
#endif // OPENCV_CUDA_WARP_SHUFFLE_HPP
|