1 /*************************************************************************** 2 smoothPolyDnans.hpp -- smooth for n>2 Dims 3 ------------------- 4 begin : May 30 2017 5 copyright : (C) 2017 by G. Duvert 6 email : m_schellens@users.sourceforge.net 7 ***************************************************************************/ 8 9 /*************************************************************************** 10 * * 11 * This program is free software; you can redistribute it and/or modify * 12 * it under the terms of the GNU General Public License as published by * 13 * the Free Software Foundation; either version 2 of the License, or * 14 * (at your option) any later version. * 15 * * 16 ***************************************************************************/ 17 18 // to be included from datatypes.cpp 19 #ifdef INCLUDE_SMOOTH_POLYD 20 21 SMOOTH_Ty* src=srcIn; 22 SMOOTH_Ty* dest=destIn; 23 SMOOTH_Ty* store; 24 25 SizeT srcDim[MAXRANK]; 26 for (int i = 0; i < rank; ++i) srcDim[i] = datainDim[i]; 27 28 SizeT nEl=1; 29 for (int i=0; i< rank; ++i) nEl*=srcDim[i]; 30 31 DUInt perm[rank]; //turntable transposition index list 32 for (int i = 0; i < rank; ++i) perm[i] = ((i + 1) % rank); //[1,2,...,0] 33 34 SizeT destStride[ MAXRANK + 1]; 35 36 SizeT destDim[ MAXRANK]; //use destDim mainly to turn resDim each time 37 for (int i = 0; i < rank; ++i) destDim[i] = srcDim[i]; 38 39 // successively apply smooth 1d and write perm-transposed values in res, then exchange res and data for next iteration. 40 // the trick is to update srcDim and resStride each time. 41 for (int r = 0; r < rank; ++r) { 42 //accelerator: compute destStride from a perm-uted srcDim: 43 destStride[0] = 1; 44 destStride[1] = srcDim[perm[0]]; 45 int m = 1; 46 for (; m < rank; ++m) destStride[m + 1] = destStride[m] * srcDim[perm[m]]; 47 for (; m < MAXRANK; ++m) destStride[m + 1] = destStride[rank]; 48 49 SizeT dimx = srcDim[0]; //which has been permutated 50 SizeT dimy = nEl / dimx; 51 SizeT w = width[r] / 2; 52 if (w == 0) {//fast transpose 53 #pragma omp parallel if (nEl >= CpuTPOOL_MIN_ELTS && (CpuTPOOL_MAX_ELTS == 0 || CpuTPOOL_MAX_ELTS <= nEl)) 54 { 55 #pragma omp for nowait 56 for (SizeT i = 0; i < nEl; ++i) dest[transposed1Index(i, srcDim, destStride, rank)] = src[i]; 57 } 58 } else { //smooth & transpose 59 #pragma omp parallel if (nEl >= CpuTPOOL_MIN_ELTS && (CpuTPOOL_MAX_ELTS == 0 || CpuTPOOL_MAX_ELTS <= nEl)) 60 { 61 #pragma omp for nowait 62 63 for (SizeT j = 0; j < dimy; ++j) { 64 //initiate mean of Width first values: 65 DDouble z; 66 DDouble n = 0; 67 DDouble mean = 0; 68 for (SizeT i = j * dimx; i < j * dimx + (2 * w + 1); ++i) { 69 DDouble v = src[i]; 70 n += 1.0; 71 z = 1. / n; 72 mean = (1. - z) * mean + z * v; 73 } 74 75 #if defined(USE_EDGE) 76 //start: use mean1 77 DDouble mean1 = mean; 78 DDouble n1 = n; 79 for (SizeT i = 0; i < w; ++i) { 80 dest[transposed1Index(w - i + j * dimx, srcDim, destStride, rank)] = mean1; 81 DDouble v = src[2 * w - i + j * dimx]; 82 mean1 -= z*v; 83 84 #if defined (EDGE_WRAP) 85 v = src[dimx - i - 1 + j * dimx]; 86 #elif defined (EDGE_TRUNCATE) 87 v = src[0 + j * dimx]; 88 #elif defined (EDGE_MIRROR) 89 v = src[i + j * dimx]; 90 #elif defined (EDGE_ZERO) 91 v = 0; 92 #endif 93 mean1 += z*v; 94 } 95 dest[transposed1Index(0 + j*dimx, srcDim, destStride, rank)] = mean1; 96 #else //just transpose 97 for (SizeT i = j * dimx; i < j * dimx + w; ++i) dest[transposed1Index(i, srcDim, destStride, rank)] = src[i]; 98 #endif 99 100 //center 101 for (SizeT i = w + j * dimx, ibef = j * dimx, iend = 2 * w + 1 + j * dimx; i < dimx - w - 1 + j * dimx; ++i, ++ibef, ++iend) { 102 dest[transposed1Index(i, srcDim, destStride, rank)] = mean; 103 DDouble v = src[ibef]; 104 mean -= z*v; 105 v = src[iend]; 106 mean += z*v; 107 } 108 dest[transposed1Index(dimx - 1 - w + j*dimx, srcDim, destStride, rank)] = mean; 109 110 #if defined(USE_EDGE) 111 //end: use mean 112 for (SizeT i = dimx - 1 - w; i < dimx - 1; ++i) { 113 dest[transposed1Index(i + j*dimx, srcDim, destStride, rank)] = mean; 114 DDouble v = src[i - w + j * dimx]; 115 mean -= z*v; 116 117 #if defined (EDGE_WRAP) 118 v = src[i - dimx + w + 1 + j * dimx]; 119 #elif defined (EDGE_TRUNCATE) 120 v = src[dimx - 1 + j * dimx]; 121 #elif defined (EDGE_MIRROR) 122 v = src[2 * dimx - i - w - 2 + j * dimx]; 123 #elif defined (EDGE_ZERO) 124 v = 0; 125 #endif 126 mean += z*v; 127 } 128 dest[transposed1Index(dimx - 1 + j*dimx, srcDim, destStride, rank)] = mean; 129 #else //just transpose 130 for (SizeT i = (dimx - w + j * dimx); i < (dimx + j * dimx); ++i) dest[transposed1Index(i, srcDim, destStride, rank)] = src[i]; 131 #endif 132 } 133 } 134 } 135 //pseudo-dim of src is now rotated by 1 136 SizeT tempSize[MAXRANK]; 137 for (int i = 0; i < rank; ++i) tempSize[i] = srcDim[i]; 138 for (int i = 0; i < rank; ++i) srcDim[i] = tempSize[perm[i]]; 139 // exchange dest and src 140 store=src; 141 src=dest; 142 dest=store; //now 'src' is the last computed array, whatever srcIn or destIn... 143 } 144 //if rank is even, must copy back src to destIn (since only the pointers were exchanged) 145 if (rank%2 == 0) memcpy(destIn, src, nEl * sizeof (src[0])); 146 #endif 147