Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 2 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_cmplx_mult_real_q31.c |
||
| 4 | * Description: Q31 complex by real multiplication |
||
| 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 groupCmplxMath |
||
| 33 | */ |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @addtogroup CmplxByRealMult |
||
| 37 | * @{ |
||
| 38 | */ |
||
| 39 | |||
| 40 | |||
| 41 | /** |
||
| 42 | * @brief Q31 complex-by-real multiplication |
||
| 43 | * @param[in] *pSrcCmplx points to the complex input vector |
||
| 44 | * @param[in] *pSrcReal points to the real input vector |
||
| 45 | * @param[out] *pCmplxDst points to the complex output vector |
||
| 46 | * @param[in] numSamples number of samples in each vector |
||
| 47 | * @return none. |
||
| 48 | * |
||
| 49 | * <b>Scaling and Overflow Behavior:</b> |
||
| 50 | * \par |
||
| 51 | * The function uses saturating arithmetic. |
||
| 52 | * Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] will be saturated. |
||
| 53 | */ |
||
| 54 | |||
| 55 | void arm_cmplx_mult_real_q31( |
||
| 56 | q31_t * pSrcCmplx, |
||
| 57 | q31_t * pSrcReal, |
||
| 58 | q31_t * pCmplxDst, |
||
| 59 | uint32_t numSamples) |
||
| 60 | { |
||
| 61 | q31_t inA1; /* Temporary variable to store input value */ |
||
| 62 | |||
| 63 | #if defined (ARM_MATH_DSP) |
||
| 64 | |||
| 65 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 66 | uint32_t blkCnt; /* loop counters */ |
||
| 67 | q31_t inA2, inA3, inA4; /* Temporary variables to hold input data */ |
||
| 68 | q31_t inB1, inB2; /* Temporary variabels to hold input data */ |
||
| 69 | q31_t out1, out2, out3, out4; /* Temporary variables to hold output data */ |
||
| 70 | |||
| 71 | /* loop Unrolling */ |
||
| 72 | blkCnt = numSamples >> 2U; |
||
| 73 | |||
| 74 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 75 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 76 | while (blkCnt > 0U) |
||
| 77 | { |
||
| 78 | /* C[2 * i] = A[2 * i] * B[i]. */ |
||
| 79 | /* C[2 * i + 1] = A[2 * i + 1] * B[i]. */ |
||
| 80 | /* read real input from complex input buffer */ |
||
| 81 | inA1 = *pSrcCmplx++; |
||
| 82 | inA2 = *pSrcCmplx++; |
||
| 83 | /* read input from real input bufer */ |
||
| 84 | inB1 = *pSrcReal++; |
||
| 85 | inB2 = *pSrcReal++; |
||
| 86 | /* read imaginary input from complex input buffer */ |
||
| 87 | inA3 = *pSrcCmplx++; |
||
| 88 | inA4 = *pSrcCmplx++; |
||
| 89 | |||
| 90 | /* multiply complex input with real input */ |
||
| 91 | out1 = ((q63_t) inA1 * inB1) >> 32; |
||
| 92 | out2 = ((q63_t) inA2 * inB1) >> 32; |
||
| 93 | out3 = ((q63_t) inA3 * inB2) >> 32; |
||
| 94 | out4 = ((q63_t) inA4 * inB2) >> 32; |
||
| 95 | |||
| 96 | /* sature the result */ |
||
| 97 | out1 = __SSAT(out1, 31); |
||
| 98 | out2 = __SSAT(out2, 31); |
||
| 99 | out3 = __SSAT(out3, 31); |
||
| 100 | out4 = __SSAT(out4, 31); |
||
| 101 | |||
| 102 | /* get result in 1.31 format */ |
||
| 103 | out1 = out1 << 1; |
||
| 104 | out2 = out2 << 1; |
||
| 105 | out3 = out3 << 1; |
||
| 106 | out4 = out4 << 1; |
||
| 107 | |||
| 108 | /* store the result to destination buffer */ |
||
| 109 | *pCmplxDst++ = out1; |
||
| 110 | *pCmplxDst++ = out2; |
||
| 111 | *pCmplxDst++ = out3; |
||
| 112 | *pCmplxDst++ = out4; |
||
| 113 | |||
| 114 | /* read real input from complex input buffer */ |
||
| 115 | inA1 = *pSrcCmplx++; |
||
| 116 | inA2 = *pSrcCmplx++; |
||
| 117 | /* read input from real input bufer */ |
||
| 118 | inB1 = *pSrcReal++; |
||
| 119 | inB2 = *pSrcReal++; |
||
| 120 | /* read imaginary input from complex input buffer */ |
||
| 121 | inA3 = *pSrcCmplx++; |
||
| 122 | inA4 = *pSrcCmplx++; |
||
| 123 | |||
| 124 | /* multiply complex input with real input */ |
||
| 125 | out1 = ((q63_t) inA1 * inB1) >> 32; |
||
| 126 | out2 = ((q63_t) inA2 * inB1) >> 32; |
||
| 127 | out3 = ((q63_t) inA3 * inB2) >> 32; |
||
| 128 | out4 = ((q63_t) inA4 * inB2) >> 32; |
||
| 129 | |||
| 130 | /* sature the result */ |
||
| 131 | out1 = __SSAT(out1, 31); |
||
| 132 | out2 = __SSAT(out2, 31); |
||
| 133 | out3 = __SSAT(out3, 31); |
||
| 134 | out4 = __SSAT(out4, 31); |
||
| 135 | |||
| 136 | /* get result in 1.31 format */ |
||
| 137 | out1 = out1 << 1; |
||
| 138 | out2 = out2 << 1; |
||
| 139 | out3 = out3 << 1; |
||
| 140 | out4 = out4 << 1; |
||
| 141 | |||
| 142 | /* store the result to destination buffer */ |
||
| 143 | *pCmplxDst++ = out1; |
||
| 144 | *pCmplxDst++ = out2; |
||
| 145 | *pCmplxDst++ = out3; |
||
| 146 | *pCmplxDst++ = out4; |
||
| 147 | |||
| 148 | /* Decrement the numSamples loop counter */ |
||
| 149 | blkCnt--; |
||
| 150 | } |
||
| 151 | |||
| 152 | /* If the numSamples is not a multiple of 4, compute any remaining output samples here. |
||
| 153 | ** No loop unrolling is used. */ |
||
| 154 | blkCnt = numSamples % 0x4U; |
||
| 155 | |||
| 156 | while (blkCnt > 0U) |
||
| 157 | { |
||
| 158 | /* C[2 * i] = A[2 * i] * B[i]. */ |
||
| 159 | /* C[2 * i + 1] = A[2 * i + 1] * B[i]. */ |
||
| 160 | /* read real input from complex input buffer */ |
||
| 161 | inA1 = *pSrcCmplx++; |
||
| 162 | inA2 = *pSrcCmplx++; |
||
| 163 | /* read input from real input bufer */ |
||
| 164 | inB1 = *pSrcReal++; |
||
| 165 | |||
| 166 | /* multiply complex input with real input */ |
||
| 167 | out1 = ((q63_t) inA1 * inB1) >> 32; |
||
| 168 | out2 = ((q63_t) inA2 * inB1) >> 32; |
||
| 169 | |||
| 170 | /* sature the result */ |
||
| 171 | out1 = __SSAT(out1, 31); |
||
| 172 | out2 = __SSAT(out2, 31); |
||
| 173 | |||
| 174 | /* get result in 1.31 format */ |
||
| 175 | out1 = out1 << 1; |
||
| 176 | out2 = out2 << 1; |
||
| 177 | |||
| 178 | /* store the result to destination buffer */ |
||
| 179 | *pCmplxDst++ = out1; |
||
| 180 | *pCmplxDst++ = out2; |
||
| 181 | |||
| 182 | /* Decrement the numSamples loop counter */ |
||
| 183 | blkCnt--; |
||
| 184 | } |
||
| 185 | |||
| 186 | #else |
||
| 187 | |||
| 188 | /* Run the below code for Cortex-M0 */ |
||
| 189 | |||
| 190 | while (numSamples > 0U) |
||
| 191 | { |
||
| 192 | /* realOut = realA * realB. */ |
||
| 193 | /* imagReal = imagA * realB. */ |
||
| 194 | inA1 = *pSrcReal++; |
||
| 195 | /* store the result in the destination buffer. */ |
||
| 196 | *pCmplxDst++ = |
||
| 197 | (q31_t) clip_q63_to_q31(((q63_t) * pSrcCmplx++ * inA1) >> 31); |
||
| 198 | *pCmplxDst++ = |
||
| 199 | (q31_t) clip_q63_to_q31(((q63_t) * pSrcCmplx++ * inA1) >> 31); |
||
| 200 | |||
| 201 | /* Decrement the numSamples loop counter */ |
||
| 202 | numSamples--; |
||
| 203 | } |
||
| 204 | |||
| 205 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 206 | |||
| 207 | } |
||
| 208 | |||
| 209 | /** |
||
| 210 | * @} end of CmplxByRealMult group |
||
| 211 | */ |