Go to most recent revision | Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 2 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_mat_sub_q31.c |
||
| 4 | * Description: Q31 matrix subtraction |
||
| 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 | #include "arm_math.h" |
||
| 30 | |||
| 31 | /** |
||
| 32 | * @ingroup groupMatrix |
||
| 33 | */ |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @addtogroup MatrixSub |
||
| 37 | * @{ |
||
| 38 | */ |
||
| 39 | |||
| 40 | /** |
||
| 41 | * @brief Q31 matrix subtraction. |
||
| 42 | * @param[in] *pSrcA points to the first input matrix structure |
||
| 43 | * @param[in] *pSrcB points to the second input matrix structure |
||
| 44 | * @param[out] *pDst points to output matrix structure |
||
| 45 | * @return The function returns either |
||
| 46 | * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking. |
||
| 47 | * |
||
| 48 | * <b>Scaling and Overflow Behavior:</b> |
||
| 49 | * \par |
||
| 50 | * The function uses saturating arithmetic. |
||
| 51 | * Results outside of the allowable Q31 range [0x80000000 0x7FFFFFFF] will be saturated. |
||
| 52 | */ |
||
| 53 | |||
| 54 | |||
| 55 | arm_status arm_mat_sub_q31( |
||
| 56 | const arm_matrix_instance_q31 * pSrcA, |
||
| 57 | const arm_matrix_instance_q31 * pSrcB, |
||
| 58 | arm_matrix_instance_q31 * pDst) |
||
| 59 | { |
||
| 60 | q31_t *pIn1 = pSrcA->pData; /* input data matrix pointer A */ |
||
| 61 | q31_t *pIn2 = pSrcB->pData; /* input data matrix pointer B */ |
||
| 62 | q31_t *pOut = pDst->pData; /* output data matrix pointer */ |
||
| 63 | q31_t inA1, inB1; /* temporary variables */ |
||
| 64 | |||
| 65 | #if defined (ARM_MATH_DSP) |
||
| 66 | |||
| 67 | q31_t inA2, inB2; /* temporary variables */ |
||
| 68 | q31_t out1, out2; /* temporary variables */ |
||
| 69 | |||
| 70 | #endif // #if defined (ARM_MATH_DSP) |
||
| 71 | |||
| 72 | uint32_t numSamples; /* total number of elements in the matrix */ |
||
| 73 | uint32_t blkCnt; /* loop counters */ |
||
| 74 | arm_status status; /* status of matrix subtraction */ |
||
| 75 | |||
| 76 | |||
| 77 | #ifdef ARM_MATH_MATRIX_CHECK |
||
| 78 | /* Check for matrix mismatch condition */ |
||
| 79 | if ((pSrcA->numRows != pSrcB->numRows) || |
||
| 80 | (pSrcA->numCols != pSrcB->numCols) || |
||
| 81 | (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols)) |
||
| 82 | { |
||
| 83 | /* Set status as ARM_MATH_SIZE_MISMATCH */ |
||
| 84 | status = ARM_MATH_SIZE_MISMATCH; |
||
| 85 | } |
||
| 86 | else |
||
| 87 | #endif |
||
| 88 | { |
||
| 89 | /* Total number of samples in the input matrix */ |
||
| 90 | numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols; |
||
| 91 | |||
| 92 | #if defined (ARM_MATH_DSP) |
||
| 93 | |||
| 94 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 95 | |||
| 96 | /* Loop Unrolling */ |
||
| 97 | blkCnt = numSamples >> 2U; |
||
| 98 | |||
| 99 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 100 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 101 | while (blkCnt > 0U) |
||
| 102 | { |
||
| 103 | /* C(m,n) = A(m,n) - B(m,n) */ |
||
| 104 | /* Subtract, saturate and then store the results in the destination buffer. */ |
||
| 105 | /* Read values from source A */ |
||
| 106 | inA1 = pIn1[0]; |
||
| 107 | |||
| 108 | /* Read values from source B */ |
||
| 109 | inB1 = pIn2[0]; |
||
| 110 | |||
| 111 | /* Read values from source A */ |
||
| 112 | inA2 = pIn1[1]; |
||
| 113 | |||
| 114 | /* Subtract and saturate */ |
||
| 115 | out1 = __QSUB(inA1, inB1); |
||
| 116 | |||
| 117 | /* Read values from source B */ |
||
| 118 | inB2 = pIn2[1]; |
||
| 119 | |||
| 120 | /* Read values from source A */ |
||
| 121 | inA1 = pIn1[2]; |
||
| 122 | |||
| 123 | /* Subtract and saturate */ |
||
| 124 | out2 = __QSUB(inA2, inB2); |
||
| 125 | |||
| 126 | /* Read values from source B */ |
||
| 127 | inB1 = pIn2[2]; |
||
| 128 | |||
| 129 | /* Store result in destination */ |
||
| 130 | pOut[0] = out1; |
||
| 131 | pOut[1] = out2; |
||
| 132 | |||
| 133 | /* Read values from source A */ |
||
| 134 | inA2 = pIn1[3]; |
||
| 135 | |||
| 136 | /* Read values from source B */ |
||
| 137 | inB2 = pIn2[3]; |
||
| 138 | |||
| 139 | /* Subtract and saturate */ |
||
| 140 | out1 = __QSUB(inA1, inB1); |
||
| 141 | |||
| 142 | /* Subtract and saturate */ |
||
| 143 | out2 = __QSUB(inA2, inB2); |
||
| 144 | |||
| 145 | /* Store result in destination */ |
||
| 146 | pOut[2] = out1; |
||
| 147 | pOut[3] = out2; |
||
| 148 | |||
| 149 | /* update pointers to process next samples */ |
||
| 150 | pIn1 += 4U; |
||
| 151 | pIn2 += 4U; |
||
| 152 | pOut += 4U; |
||
| 153 | |||
| 154 | /* Decrement the loop counter */ |
||
| 155 | blkCnt--; |
||
| 156 | } |
||
| 157 | |||
| 158 | /* If the numSamples is not a multiple of 4, compute any remaining output samples here. |
||
| 159 | ** No loop unrolling is used. */ |
||
| 160 | blkCnt = numSamples % 0x4U; |
||
| 161 | |||
| 162 | #else |
||
| 163 | |||
| 164 | /* Run the below code for Cortex-M0 */ |
||
| 165 | |||
| 166 | /* Initialize blkCnt with number of samples */ |
||
| 167 | blkCnt = numSamples; |
||
| 168 | |||
| 169 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 170 | |||
| 171 | while (blkCnt > 0U) |
||
| 172 | { |
||
| 173 | /* C(m,n) = A(m,n) - B(m,n) */ |
||
| 174 | /* Subtract, saturate and then store the results in the destination buffer. */ |
||
| 175 | inA1 = *pIn1++; |
||
| 176 | inB1 = *pIn2++; |
||
| 177 | |||
| 178 | inA1 = __QSUB(inA1, inB1); |
||
| 179 | |||
| 180 | *pOut++ = inA1; |
||
| 181 | |||
| 182 | /* Decrement the loop counter */ |
||
| 183 | blkCnt--; |
||
| 184 | } |
||
| 185 | |||
| 186 | /* Set status as ARM_MATH_SUCCESS */ |
||
| 187 | status = ARM_MATH_SUCCESS; |
||
| 188 | } |
||
| 189 | |||
| 190 | /* Return to application */ |
||
| 191 | return (status); |
||
| 192 | } |
||
| 193 | |||
| 194 | /** |
||
| 195 | * @} end of MatrixSub group |
||
| 196 | */ |