| 1 | /* ----------------------------------------------------------------------
|
|---|
| 2 | * Project: CMSIS DSP Library
|
|---|
| 3 | * Title: arm_power_q15.c
|
|---|
| 4 | * Description: Sum of the squares of the elements of a Q15 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 Q15 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 64-bit internal accumulator.
|
|---|
| 52 | * The input is represented in 1.15 format.
|
|---|
| 53 | * Intermediate multiplication yields a 2.30 format, and this
|
|---|
| 54 | * result is added without saturation to a 64-bit accumulator in 34.30 format.
|
|---|
| 55 | * With 33 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 34.30 format.
|
|---|
| 58 | *
|
|---|
| 59 | */
|
|---|
| 60 |
|
|---|
| 61 | void arm_power_q15(
|
|---|
| 62 | q15_t * pSrc,
|
|---|
| 63 | uint32_t blockSize,
|
|---|
| 64 | q63_t * pResult)
|
|---|
| 65 | {
|
|---|
| 66 | q63_t sum = 0; /* Temporary result storage */
|
|---|
| 67 |
|
|---|
| 68 | #if defined (ARM_MATH_DSP)
|
|---|
| 69 | /* Run the below code for Cortex-M4 and Cortex-M3 */
|
|---|
| 70 |
|
|---|
| 71 | q31_t in32; /* Temporary variable to store input value */
|
|---|
| 72 | q15_t in16; /* Temporary variable to store input value */
|
|---|
| 73 | uint32_t blkCnt; /* loop counter */
|
|---|
| 74 |
|
|---|
| 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 | /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
|
|---|
| 84 | /* Compute Power and then store the result in a temporary variable, sum. */
|
|---|
| 85 | in32 = *__SIMD32(pSrc)++;
|
|---|
| 86 | sum = __SMLALD(in32, in32, sum);
|
|---|
| 87 | in32 = *__SIMD32(pSrc)++;
|
|---|
| 88 | sum = __SMLALD(in32, in32, sum);
|
|---|
| 89 |
|
|---|
| 90 | /* Decrement the loop counter */
|
|---|
| 91 | blkCnt--;
|
|---|
| 92 | }
|
|---|
| 93 |
|
|---|
| 94 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
|
|---|
| 95 | ** No loop unrolling is used. */
|
|---|
| 96 | blkCnt = blockSize % 0x4U;
|
|---|
| 97 |
|
|---|
| 98 | while (blkCnt > 0U)
|
|---|
| 99 | {
|
|---|
| 100 | /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
|
|---|
| 101 | /* Compute Power and then store the result in a temporary variable, sum. */
|
|---|
| 102 | in16 = *pSrc++;
|
|---|
| 103 | sum = __SMLALD(in16, in16, sum);
|
|---|
| 104 |
|
|---|
| 105 | /* Decrement the loop counter */
|
|---|
| 106 | blkCnt--;
|
|---|
| 107 | }
|
|---|
| 108 |
|
|---|
| 109 | #else
|
|---|
| 110 | /* Run the below code for Cortex-M0 */
|
|---|
| 111 |
|
|---|
| 112 | q15_t in; /* Temporary variable to store input value */
|
|---|
| 113 | uint32_t blkCnt; /* loop counter */
|
|---|
| 114 |
|
|---|
| 115 |
|
|---|
| 116 | /* Loop over blockSize number of values */
|
|---|
| 117 | blkCnt = blockSize;
|
|---|
| 118 |
|
|---|
| 119 | while (blkCnt > 0U)
|
|---|
| 120 | {
|
|---|
| 121 | /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
|
|---|
| 122 | /* Compute Power and then store the result in a temporary variable, sum. */
|
|---|
| 123 | in = *pSrc++;
|
|---|
| 124 | sum += ((q31_t) in * in);
|
|---|
| 125 |
|
|---|
| 126 | /* Decrement the loop counter */
|
|---|
| 127 | blkCnt--;
|
|---|
| 128 | }
|
|---|
| 129 |
|
|---|
| 130 | #endif /* #if defined (ARM_MATH_DSP) */
|
|---|
| 131 |
|
|---|
| 132 | /* Store the results in 34.30 format */
|
|---|
| 133 | *pResult = sum;
|
|---|
| 134 | }
|
|---|
| 135 |
|
|---|
| 136 | /**
|
|---|
| 137 | * @} end of power group
|
|---|
| 138 | */
|
|---|
