Rev 56 | Details | Compare with Previous | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 56 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_mult_f32.c |
||
| 4 | * Description: Floating-point vector 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 groupMath |
||
| 33 | */ |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @defgroup BasicMult Vector Multiplication |
||
| 37 | * |
||
| 38 | * Element-by-element multiplication of two vectors. |
||
| 39 | * |
||
| 40 | * <pre> |
||
| 41 | * pDst[n] = pSrcA[n] * pSrcB[n], 0 <= n < blockSize. |
||
| 42 | * </pre> |
||
| 43 | * |
||
| 44 | * There are separate functions for floating-point, Q7, Q15, and Q31 data types. |
||
| 45 | */ |
||
| 46 | |||
| 47 | /** |
||
| 48 | * @addtogroup BasicMult |
||
| 49 | * @{ |
||
| 50 | */ |
||
| 51 | |||
| 52 | /** |
||
| 53 | * @brief Floating-point vector multiplication. |
||
| 54 | * @param[in] *pSrcA points to the first input vector |
||
| 55 | * @param[in] *pSrcB points to the second input vector |
||
| 56 | * @param[out] *pDst points to the output vector |
||
| 57 | * @param[in] blockSize number of samples in each vector |
||
| 58 | * @return none. |
||
| 59 | */ |
||
| 60 | |||
| 61 | void arm_mult_f32( |
||
| 62 | float32_t * pSrcA, |
||
| 63 | float32_t * pSrcB, |
||
| 64 | float32_t * pDst, |
||
| 65 | uint32_t blockSize) |
||
| 66 | { |
||
| 67 | uint32_t blkCnt; /* loop counters */ |
||
| 68 | #if defined (ARM_MATH_DSP) |
||
| 69 | |||
| 70 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 71 | float32_t inA1, inA2, inA3, inA4; /* temporary input variables */ |
||
| 72 | float32_t inB1, inB2, inB3, inB4; /* temporary input variables */ |
||
| 73 | float32_t out1, out2, out3, out4; /* temporary output variables */ |
||
| 74 | |||
| 75 | /* loop Unrolling */ |
||
| 76 | blkCnt = blockSize >> 2U; |
||
| 77 | |||
| 78 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 79 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 80 | while (blkCnt > 0U) |
||
| 81 | { |
||
| 82 | /* C = A * B */ |
||
| 83 | /* Multiply the inputs and store the results in output buffer */ |
||
| 84 | /* read sample from sourceA */ |
||
| 85 | inA1 = *pSrcA; |
||
| 86 | /* read sample from sourceB */ |
||
| 87 | inB1 = *pSrcB; |
||
| 88 | /* read sample from sourceA */ |
||
| 89 | inA2 = *(pSrcA + 1); |
||
| 90 | /* read sample from sourceB */ |
||
| 91 | inB2 = *(pSrcB + 1); |
||
| 92 | |||
| 93 | /* out = sourceA * sourceB */ |
||
| 94 | out1 = inA1 * inB1; |
||
| 95 | |||
| 96 | /* read sample from sourceA */ |
||
| 97 | inA3 = *(pSrcA + 2); |
||
| 98 | /* read sample from sourceB */ |
||
| 99 | inB3 = *(pSrcB + 2); |
||
| 100 | |||
| 101 | /* out = sourceA * sourceB */ |
||
| 102 | out2 = inA2 * inB2; |
||
| 103 | |||
| 104 | /* read sample from sourceA */ |
||
| 105 | inA4 = *(pSrcA + 3); |
||
| 106 | |||
| 107 | /* store result to destination buffer */ |
||
| 108 | *pDst = out1; |
||
| 109 | |||
| 110 | /* read sample from sourceB */ |
||
| 111 | inB4 = *(pSrcB + 3); |
||
| 112 | |||
| 113 | /* out = sourceA * sourceB */ |
||
| 114 | out3 = inA3 * inB3; |
||
| 115 | |||
| 116 | /* store result to destination buffer */ |
||
| 117 | *(pDst + 1) = out2; |
||
| 118 | |||
| 119 | /* out = sourceA * sourceB */ |
||
| 120 | out4 = inA4 * inB4; |
||
| 121 | /* store result to destination buffer */ |
||
| 122 | *(pDst + 2) = out3; |
||
| 123 | /* store result to destination buffer */ |
||
| 124 | *(pDst + 3) = out4; |
||
| 125 | |||
| 126 | |||
| 127 | /* update pointers to process next samples */ |
||
| 128 | pSrcA += 4U; |
||
| 129 | pSrcB += 4U; |
||
| 130 | pDst += 4U; |
||
| 131 | |||
| 132 | /* Decrement the blockSize loop counter */ |
||
| 133 | blkCnt--; |
||
| 134 | } |
||
| 135 | |||
| 136 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here. |
||
| 137 | ** No loop unrolling is used. */ |
||
| 138 | blkCnt = blockSize % 0x4U; |
||
| 139 | |||
| 140 | #else |
||
| 141 | |||
| 142 | /* Run the below code for Cortex-M0 */ |
||
| 143 | |||
| 144 | /* Initialize blkCnt with number of samples */ |
||
| 145 | blkCnt = blockSize; |
||
| 146 | |||
| 147 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 148 | |||
| 149 | while (blkCnt > 0U) |
||
| 150 | { |
||
| 151 | /* C = A * B */ |
||
| 152 | /* Multiply the inputs and store the results in output buffer */ |
||
| 153 | *pDst++ = (*pSrcA++) * (*pSrcB++); |
||
| 154 | |||
| 155 | /* Decrement the blockSize loop counter */ |
||
| 156 | blkCnt--; |
||
| 157 | } |
||
| 158 | } |
||
| 159 | |||
| 160 | /** |
||
| 161 | * @} end of BasicMult group |
||
| 162 | */ |