Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 56 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_mat_trans_f32.c |
||
| 4 | * Description: Floating-point matrix transpose |
||
| 5 | * |
||
| 6 | * $Date: 27. January 2017 |
||
| 7 | * $Revision: V.1.5.1 |
||
| 8 | * |
||
| 9 | * Target Processor: Cortex-M cores |
||
| 10 | * -------------------------------------------------------------------- */ |
||
| 11 | /* |
||
| 12 | * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved. |
||
| 13 | * |
||
| 14 | * SPDX-License-Identifier: Apache-2.0 |
||
| 15 | * |
||
| 16 | * Licensed under the Apache License, Version 2.0 (the License); you may |
||
| 17 | * not use this file except in compliance with the License. |
||
| 18 | * You may obtain a copy of the License at |
||
| 19 | * |
||
| 20 | * www.apache.org/licenses/LICENSE-2.0 |
||
| 21 | * |
||
| 22 | * Unless required by applicable law or agreed to in writing, software |
||
| 23 | * distributed under the License is distributed on an AS IS BASIS, WITHOUT |
||
| 24 | * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
||
| 25 | * See the License for the specific language governing permissions and |
||
| 26 | * limitations under the License. |
||
| 27 | */ |
||
| 28 | |||
| 29 | /** |
||
| 30 | * @defgroup MatrixTrans Matrix Transpose |
||
| 31 | * |
||
| 32 | * Tranposes a matrix. |
||
| 33 | * Transposing an <code>M x N</code> matrix flips it around the center diagonal and results in an <code>N x M</code> matrix. |
||
| 34 | * \image html MatrixTranspose.gif "Transpose of a 3 x 3 matrix" |
||
| 35 | */ |
||
| 36 | |||
| 37 | #include "arm_math.h" |
||
| 38 | |||
| 39 | /** |
||
| 40 | * @ingroup groupMatrix |
||
| 41 | */ |
||
| 42 | |||
| 43 | /** |
||
| 44 | * @addtogroup MatrixTrans |
||
| 45 | * @{ |
||
| 46 | */ |
||
| 47 | |||
| 48 | /** |
||
| 49 | * @brief Floating-point matrix transpose. |
||
| 50 | * @param[in] *pSrc points to the input matrix |
||
| 51 | * @param[out] *pDst points to the output matrix |
||
| 52 | * @return The function returns either <code>ARM_MATH_SIZE_MISMATCH</code> |
||
| 53 | * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking. |
||
| 54 | */ |
||
| 55 | |||
| 56 | |||
| 57 | arm_status arm_mat_trans_f32( |
||
| 58 | const arm_matrix_instance_f32 * pSrc, |
||
| 59 | arm_matrix_instance_f32 * pDst) |
||
| 60 | { |
||
| 61 | float32_t *pIn = pSrc->pData; /* input data matrix pointer */ |
||
| 62 | float32_t *pOut = pDst->pData; /* output data matrix pointer */ |
||
| 63 | float32_t *px; /* Temporary output data matrix pointer */ |
||
| 64 | uint16_t nRows = pSrc->numRows; /* number of rows */ |
||
| 65 | uint16_t nColumns = pSrc->numCols; /* number of columns */ |
||
| 66 | |||
| 67 | #if defined (ARM_MATH_DSP) |
||
| 68 | |||
| 69 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 70 | |||
| 71 | uint16_t blkCnt, i = 0U, row = nRows; /* loop counters */ |
||
| 72 | arm_status status; /* status of matrix transpose */ |
||
| 73 | |||
| 74 | |||
| 75 | #ifdef ARM_MATH_MATRIX_CHECK |
||
| 76 | |||
| 77 | |||
| 78 | /* Check for matrix mismatch condition */ |
||
| 79 | if ((pSrc->numRows != pDst->numCols) || (pSrc->numCols != pDst->numRows)) |
||
| 80 | { |
||
| 81 | /* Set status as ARM_MATH_SIZE_MISMATCH */ |
||
| 82 | status = ARM_MATH_SIZE_MISMATCH; |
||
| 83 | } |
||
| 84 | else |
||
| 85 | #endif /* #ifdef ARM_MATH_MATRIX_CHECK */ |
||
| 86 | |||
| 87 | { |
||
| 88 | /* Matrix transpose by exchanging the rows with columns */ |
||
| 89 | /* row loop */ |
||
| 90 | do |
||
| 91 | { |
||
| 92 | /* Loop Unrolling */ |
||
| 93 | blkCnt = nColumns >> 2; |
||
| 94 | |||
| 95 | /* The pointer px is set to starting address of the column being processed */ |
||
| 96 | px = pOut + i; |
||
| 97 | |||
| 98 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 99 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 100 | while (blkCnt > 0U) /* column loop */ |
||
| 101 | { |
||
| 102 | /* Read and store the input element in the destination */ |
||
| 103 | *px = *pIn++; |
||
| 104 | |||
| 105 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 106 | px += nRows; |
||
| 107 | |||
| 108 | /* Read and store the input element in the destination */ |
||
| 109 | *px = *pIn++; |
||
| 110 | |||
| 111 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 112 | px += nRows; |
||
| 113 | |||
| 114 | /* Read and store the input element in the destination */ |
||
| 115 | *px = *pIn++; |
||
| 116 | |||
| 117 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 118 | px += nRows; |
||
| 119 | |||
| 120 | /* Read and store the input element in the destination */ |
||
| 121 | *px = *pIn++; |
||
| 122 | |||
| 123 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 124 | px += nRows; |
||
| 125 | |||
| 126 | /* Decrement the column loop counter */ |
||
| 127 | blkCnt--; |
||
| 128 | } |
||
| 129 | |||
| 130 | /* Perform matrix transpose for last 3 samples here. */ |
||
| 131 | blkCnt = nColumns % 0x4U; |
||
| 132 | |||
| 133 | while (blkCnt > 0U) |
||
| 134 | { |
||
| 135 | /* Read and store the input element in the destination */ |
||
| 136 | *px = *pIn++; |
||
| 137 | |||
| 138 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 139 | px += nRows; |
||
| 140 | |||
| 141 | /* Decrement the column loop counter */ |
||
| 142 | blkCnt--; |
||
| 143 | } |
||
| 144 | |||
| 145 | #else |
||
| 146 | |||
| 147 | /* Run the below code for Cortex-M0 */ |
||
| 148 | |||
| 149 | uint16_t col, i = 0U, row = nRows; /* loop counters */ |
||
| 150 | arm_status status; /* status of matrix transpose */ |
||
| 151 | |||
| 152 | |||
| 153 | #ifdef ARM_MATH_MATRIX_CHECK |
||
| 154 | |||
| 155 | /* Check for matrix mismatch condition */ |
||
| 156 | if ((pSrc->numRows != pDst->numCols) || (pSrc->numCols != pDst->numRows)) |
||
| 157 | { |
||
| 158 | /* Set status as ARM_MATH_SIZE_MISMATCH */ |
||
| 159 | status = ARM_MATH_SIZE_MISMATCH; |
||
| 160 | } |
||
| 161 | else |
||
| 162 | #endif /* #ifdef ARM_MATH_MATRIX_CHECK */ |
||
| 163 | |||
| 164 | { |
||
| 165 | /* Matrix transpose by exchanging the rows with columns */ |
||
| 166 | /* row loop */ |
||
| 167 | do |
||
| 168 | { |
||
| 169 | /* The pointer px is set to starting address of the column being processed */ |
||
| 170 | px = pOut + i; |
||
| 171 | |||
| 172 | /* Initialize column loop counter */ |
||
| 173 | col = nColumns; |
||
| 174 | |||
| 175 | while (col > 0U) |
||
| 176 | { |
||
| 177 | /* Read and store the input element in the destination */ |
||
| 178 | *px = *pIn++; |
||
| 179 | |||
| 180 | /* Update the pointer px to point to the next row of the transposed matrix */ |
||
| 181 | px += nRows; |
||
| 182 | |||
| 183 | /* Decrement the column loop counter */ |
||
| 184 | col--; |
||
| 185 | } |
||
| 186 | |||
| 187 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 188 | |||
| 189 | i++; |
||
| 190 | |||
| 191 | /* Decrement the row loop counter */ |
||
| 192 | row--; |
||
| 193 | |||
| 194 | } while (row > 0U); /* row loop end */ |
||
| 195 | |||
| 196 | /* Set status as ARM_MATH_SUCCESS */ |
||
| 197 | status = ARM_MATH_SUCCESS; |
||
| 198 | } |
||
| 199 | |||
| 200 | /* Return to application */ |
||
| 201 | return (status); |
||
| 202 | } |
||
| 203 | |||
| 204 | /** |
||
| 205 | * @} end of MatrixTrans group |
||
| 206 | */ |