Lines Matching full:nthreads
590 "#define NTHREADS 256\n"
756 "const int nthreads, const int block_hist_size, const int img_block_width,\n"
769 "float sum = reduce_smem(squares, nthreads);\n"
775 "sum = reduce_smem(squares, nthreads);\n"
857 "__local float products[NTHREADS];\n"
908 "for (int i = tid; i < cdescr_size; i += NTHREADS)\n"
915 "__local float products[NTHREADS];\n"
966 "for (int i = tid; i < cdescr_size; i += NTHREADS)\n"
986 "for (int i = tid; i < cdescr_size; i += NTHREADS)\n"
1007 "__local float sh_row[(NTHREADS + 2) * 3];\n"
1014 "sh_row[tid + 1 + (NTHREADS + 2)] = val.y;\n"
1015 "sh_row[tid + 1 + 2 * (NTHREADS + 2)] = val.z;\n"
1020 "sh_row[(NTHREADS + 2)] = val.y;\n"
1021 "sh_row[2 * (NTHREADS + 2)] = val.z;\n"
1027 "sh_row[gSizeX + 1 + (NTHREADS + 2)] = val.y;\n"
1028 "sh_row[gSizeX + 1 + 2 * (NTHREADS + 2)] = val.z;\n"
1033 "float4 a = (float4) (sh_row[tid], sh_row[tid + (NTHREADS + 2)],\n"
1034 "sh_row[tid + 2 * (NTHREADS + 2)], 0);\n"
1035 "float4 b = (float4) (sh_row[tid + 2], sh_row[tid + 2 + (NTHREADS + 2)],\n"
1036 "sh_row[tid + 2 + 2 * (NTHREADS + 2)], 0);\n"
1089 "__local float sh_row[NTHREADS + 2];\n"
