Go to most recent revision | Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 56 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_power_q7.c |
||
| 4 | * Description: Sum of the squares of the elements of a Q7 vector |
||
| 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 groupStats |
||
| 33 | */ |
||
| 34 | |||
| 35 | /** |
||
| 36 | * @addtogroup power |
||
| 37 | * @{ |
||
| 38 | */ |
||
| 39 | |||
| 40 | /** |
||
| 41 | * @brief Sum of the squares of the elements of a Q7 vector. |
||
| 42 | * @param[in] *pSrc points to the input vector |
||
| 43 | * @param[in] blockSize length of the input vector |
||
| 44 | * @param[out] *pResult sum of the squares value returned here |
||
| 45 | * @return none. |
||
| 46 | * |
||
| 47 | * @details |
||
| 48 | * <b>Scaling and Overflow Behavior:</b> |
||
| 49 | * |
||
| 50 | * \par |
||
| 51 | * The function is implemented using a 32-bit internal accumulator. |
||
| 52 | * The input is represented in 1.7 format. |
||
| 53 | * Intermediate multiplication yields a 2.14 format, and this |
||
| 54 | * result is added without saturation to an accumulator in 18.14 format. |
||
| 55 | * With 17 guard bits in the accumulator, there is no risk of overflow, and the |
||
| 56 | * full precision of the intermediate multiplication is preserved. |
||
| 57 | * Finally, the return result is in 18.14 format. |
||
| 58 | * |
||
| 59 | */ |
||
| 60 | |||
| 61 | void arm_power_q7( |
||
| 62 | q7_t * pSrc, |
||
| 63 | uint32_t blockSize, |
||
| 64 | q31_t * pResult) |
||
| 65 | { |
||
| 66 | q31_t sum = 0; /* Temporary result storage */ |
||
| 67 | q7_t in; /* Temporary variable to store input */ |
||
| 68 | uint32_t blkCnt; /* loop counter */ |
||
| 69 | |||
| 70 | #if defined (ARM_MATH_DSP) |
||
| 71 | /* Run the below code for Cortex-M4 and Cortex-M3 */ |
||
| 72 | |||
| 73 | q31_t input1; /* Temporary variable to store packed input */ |
||
| 74 | q31_t in1, in2; /* Temporary variables to store input */ |
||
| 75 | |||
| 76 | /*loop Unrolling */ |
||
| 77 | blkCnt = blockSize >> 2U; |
||
| 78 | |||
| 79 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
| 80 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
| 81 | while (blkCnt > 0U) |
||
| 82 | { |
||
| 83 | /* Reading two inputs of pSrc vector and packing */ |
||
| 84 | input1 = *__SIMD32(pSrc)++; |
||
| 85 | |||
| 86 | in1 = __SXTB16(__ROR(input1, 8)); |
||
| 87 | in2 = __SXTB16(input1); |
||
| 88 | |||
| 89 | /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ |
||
| 90 | /* calculate power and accumulate to accumulator */ |
||
| 91 | sum = __SMLAD(in1, in1, sum); |
||
| 92 | sum = __SMLAD(in2, in2, sum); |
||
| 93 | |||
| 94 | /* Decrement the loop counter */ |
||
| 95 | blkCnt--; |
||
| 96 | } |
||
| 97 | |||
| 98 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here. |
||
| 99 | ** No loop unrolling is used. */ |
||
| 100 | blkCnt = blockSize % 0x4U; |
||
| 101 | |||
| 102 | #else |
||
| 103 | /* Run the below code for Cortex-M0 */ |
||
| 104 | |||
| 105 | /* Loop over blockSize number of values */ |
||
| 106 | blkCnt = blockSize; |
||
| 107 | |||
| 108 | #endif /* #if defined (ARM_MATH_DSP) */ |
||
| 109 | |||
| 110 | while (blkCnt > 0U) |
||
| 111 | { |
||
| 112 | /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ |
||
| 113 | /* Compute Power and then store the result in a temporary variable, sum. */ |
||
| 114 | in = *pSrc++; |
||
| 115 | sum += ((q15_t) in * in); |
||
| 116 | |||
| 117 | /* Decrement the loop counter */ |
||
| 118 | blkCnt--; |
||
| 119 | } |
||
| 120 | |||
| 121 | /* Store the result in 18.14 format */ |
||
| 122 | *pResult = sum; |
||
| 123 | } |
||
| 124 | |||
| 125 | /** |
||
| 126 | * @} end of power group |
||
| 127 | */ |