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_REDUCE_HPP__ 44 #define __OPENCV_CUDA_REDUCE_HPP__ 45 46 #include <thrust/tuple.h> 47 #include "detail/reduce.hpp" 48 #include "detail/reduce_key_val.hpp" 49 50 /** @file 51 * @deprecated Use @ref cudev instead. 52 */ 53 54 //! @cond IGNORED 55 56 namespace cv { namespace cuda { namespace device 57 { 58 template <int N, typename T, class Op> 59 __device__ __forceinline__ void reduce(volatile T* smem, T& val, unsigned int tid, const Op& op) 60 { 61 reduce_detail::Dispatcher<N>::reductor::template reduce<volatile T*, T&, const Op&>(smem, val, tid, op); 62 } 63 template <int N, 64 typename P0, typename P1, typename P2, typename P3, typename P4, typename P5, typename P6, typename P7, typename P8, typename P9, 65 typename R0, typename R1, typename R2, typename R3, typename R4, typename R5, typename R6, typename R7, typename R8, typename R9, 66 class Op0, class Op1, class Op2, class Op3, class Op4, class Op5, class Op6, class Op7, class Op8, class Op9> 67 __device__ __forceinline__ void reduce(const thrust::tuple<P0, P1, P2, P3, P4, P5, P6, P7, P8, P9>& smem, 68 const thrust::tuple<R0, R1, R2, R3, R4, R5, R6, R7, R8, R9>& val, 69 unsigned int tid, 70 const thrust::tuple<Op0, Op1, Op2, Op3, Op4, Op5, Op6, Op7, Op8, Op9>& op) 71 { 72 reduce_detail::Dispatcher<N>::reductor::template reduce< 73 const thrust::tuple<P0, P1, P2, P3, P4, P5, P6, P7, P8, P9>&, 74 const thrust::tuple<R0, R1, R2, R3, R4, R5, R6, R7, R8, R9>&, 75 const thrust::tuple<Op0, Op1, Op2, Op3, Op4, Op5, Op6, Op7, Op8, Op9>&>(smem, val, tid, op); 76 } 77 78 template <unsigned int N, typename K, typename V, class Cmp> 79 __device__ __forceinline__ void reduceKeyVal(volatile K* skeys, K& key, volatile V* svals, V& val, unsigned int tid, const Cmp& cmp) 80 { 81 reduce_key_val_detail::Dispatcher<N>::reductor::template reduce<volatile K*, K&, volatile V*, V&, const Cmp&>(skeys, key, svals, val, tid, cmp); 82 } 83 template <unsigned int N, 84 typename K, 85 typename VP0, typename VP1, typename VP2, typename VP3, typename VP4, typename VP5, typename VP6, typename VP7, typename VP8, typename VP9, 86 typename VR0, typename VR1, typename VR2, typename VR3, typename VR4, typename VR5, typename VR6, typename VR7, typename VR8, typename VR9, 87 class Cmp> 88 __device__ __forceinline__ void reduceKeyVal(volatile K* skeys, K& key, 89 const thrust::tuple<VP0, VP1, VP2, VP3, VP4, VP5, VP6, VP7, VP8, VP9>& svals, 90 const thrust::tuple<VR0, VR1, VR2, VR3, VR4, VR5, VR6, VR7, VR8, VR9>& val, 91 unsigned int tid, const Cmp& cmp) 92 { 93 reduce_key_val_detail::Dispatcher<N>::reductor::template reduce<volatile K*, K&, 94 const thrust::tuple<VP0, VP1, VP2, VP3, VP4, VP5, VP6, VP7, VP8, VP9>&, 95 const thrust::tuple<VR0, VR1, VR2, VR3, VR4, VR5, VR6, VR7, VR8, VR9>&, 96 const Cmp&>(skeys, key, svals, val, tid, cmp); 97 } 98 template <unsigned int N, 99 typename KP0, typename KP1, typename KP2, typename KP3, typename KP4, typename KP5, typename KP6, typename KP7, typename KP8, typename KP9, 100 typename KR0, typename KR1, typename KR2, typename KR3, typename KR4, typename KR5, typename KR6, typename KR7, typename KR8, typename KR9, 101 typename VP0, typename VP1, typename VP2, typename VP3, typename VP4, typename VP5, typename VP6, typename VP7, typename VP8, typename VP9, 102 typename VR0, typename VR1, typename VR2, typename VR3, typename VR4, typename VR5, typename VR6, typename VR7, typename VR8, typename VR9, 103 class Cmp0, class Cmp1, class Cmp2, class Cmp3, class Cmp4, class Cmp5, class Cmp6, class Cmp7, class Cmp8, class Cmp9> 104 __device__ __forceinline__ void reduceKeyVal(const thrust::tuple<KP0, KP1, KP2, KP3, KP4, KP5, KP6, KP7, KP8, KP9>& skeys, 105 const thrust::tuple<KR0, KR1, KR2, KR3, KR4, KR5, KR6, KR7, KR8, KR9>& key, 106 const thrust::tuple<VP0, VP1, VP2, VP3, VP4, VP5, VP6, VP7, VP8, VP9>& svals, 107 const thrust::tuple<VR0, VR1, VR2, VR3, VR4, VR5, VR6, VR7, VR8, VR9>& val, 108 unsigned int tid, 109 const thrust::tuple<Cmp0, Cmp1, Cmp2, Cmp3, Cmp4, Cmp5, Cmp6, Cmp7, Cmp8, Cmp9>& cmp) 110 { 111 reduce_key_val_detail::Dispatcher<N>::reductor::template reduce< 112 const thrust::tuple<KP0, KP1, KP2, KP3, KP4, KP5, KP6, KP7, KP8, KP9>&, 113 const thrust::tuple<KR0, KR1, KR2, KR3, KR4, KR5, KR6, KR7, KR8, KR9>&, 114 const thrust::tuple<VP0, VP1, VP2, VP3, VP4, VP5, VP6, VP7, VP8, VP9>&, 115 const thrust::tuple<VR0, VR1, VR2, VR3, VR4, VR5, VR6, VR7, VR8, VR9>&, 116 const thrust::tuple<Cmp0, Cmp1, Cmp2, Cmp3, Cmp4, Cmp5, Cmp6, Cmp7, Cmp8, Cmp9>& 117 >(skeys, key, svals, val, tid, cmp); 118 } 119 120 // smem_tuple 121 122 template <typename T0> 123 __device__ __forceinline__ 124 thrust::tuple<volatile T0*> 125 smem_tuple(T0* t0) 126 { 127 return thrust::make_tuple((volatile T0*) t0); 128 } 129 130 template <typename T0, typename T1> 131 __device__ __forceinline__ 132 thrust::tuple<volatile T0*, volatile T1*> 133 smem_tuple(T0* t0, T1* t1) 134 { 135 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1); 136 } 137 138 template <typename T0, typename T1, typename T2> 139 __device__ __forceinline__ 140 thrust::tuple<volatile T0*, volatile T1*, volatile T2*> 141 smem_tuple(T0* t0, T1* t1, T2* t2) 142 { 143 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2); 144 } 145 146 template <typename T0, typename T1, typename T2, typename T3> 147 __device__ __forceinline__ 148 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*> 149 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3) 150 { 151 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3); 152 } 153 154 template <typename T0, typename T1, typename T2, typename T3, typename T4> 155 __device__ __forceinline__ 156 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*> 157 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4) 158 { 159 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4); 160 } 161 162 template <typename T0, typename T1, typename T2, typename T3, typename T4, typename T5> 163 __device__ __forceinline__ 164 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*, volatile T5*> 165 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4, T5* t5) 166 { 167 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4, (volatile T5*) t5); 168 } 169 170 template <typename T0, typename T1, typename T2, typename T3, typename T4, typename T5, typename T6> 171 __device__ __forceinline__ 172 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*, volatile T5*, volatile T6*> 173 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4, T5* t5, T6* t6) 174 { 175 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4, (volatile T5*) t5, (volatile T6*) t6); 176 } 177 178 template <typename T0, typename T1, typename T2, typename T3, typename T4, typename T5, typename T6, typename T7> 179 __device__ __forceinline__ 180 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*, volatile T5*, volatile T6*, volatile T7*> 181 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4, T5* t5, T6* t6, T7* t7) 182 { 183 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4, (volatile T5*) t5, (volatile T6*) t6, (volatile T7*) t7); 184 } 185 186 template <typename T0, typename T1, typename T2, typename T3, typename T4, typename T5, typename T6, typename T7, typename T8> 187 __device__ __forceinline__ 188 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*, volatile T5*, volatile T6*, volatile T7*, volatile T8*> 189 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4, T5* t5, T6* t6, T7* t7, T8* t8) 190 { 191 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4, (volatile T5*) t5, (volatile T6*) t6, (volatile T7*) t7, (volatile T8*) t8); 192 } 193 194 template <typename T0, typename T1, typename T2, typename T3, typename T4, typename T5, typename T6, typename T7, typename T8, typename T9> 195 __device__ __forceinline__ 196 thrust::tuple<volatile T0*, volatile T1*, volatile T2*, volatile T3*, volatile T4*, volatile T5*, volatile T6*, volatile T7*, volatile T8*, volatile T9*> 197 smem_tuple(T0* t0, T1* t1, T2* t2, T3* t3, T4* t4, T5* t5, T6* t6, T7* t7, T8* t8, T9* t9) 198 { 199 return thrust::make_tuple((volatile T0*) t0, (volatile T1*) t1, (volatile T2*) t2, (volatile T3*) t3, (volatile T4*) t4, (volatile T5*) t5, (volatile T6*) t6, (volatile T7*) t7, (volatile T8*) t8, (volatile T9*) t9); 200 } 201 }}} 202 203 //! @endcond 204 205 #endif // __OPENCV_CUDA_UTILITY_HPP__ 206
