Analog Devices / Mbed OS EVAL-CN0540-ARDZ

Dependencies:   platform_drivers LTC26X6 AD77681

CMSIS-DSP_Lib/Source/StatisticsFunctions/arm_power_q15.c@1:9dd7c64b4a64, 2021-12-06 (annotated)

Committer:
jngarlitos
Date:
Mon Dec 06 05:22:28 2021 +0000
Revision:
1:9dd7c64b4a64
EVAL-CN0540-ARDZ mbed example program Initial Commit

Who changed what in which revision?

UserRevisionLine numberNew contents of line
jngarlitos 1:9dd7c64b4a64 1 /* ----------------------------------------------------------------------
jngarlitos 1:9dd7c64b4a64 2 * Copyright (C) 2010-2014 ARM Limited. All rights reserved.
jngarlitos 1:9dd7c64b4a64 3 *
jngarlitos 1:9dd7c64b4a64 4 * $Date: 19. March 2015
jngarlitos 1:9dd7c64b4a64 5 * $Revision: V.1.4.5
jngarlitos 1:9dd7c64b4a64 6 *
jngarlitos 1:9dd7c64b4a64 7 * Project: CMSIS DSP Library
jngarlitos 1:9dd7c64b4a64 8 * Title: arm_power_q15.c
jngarlitos 1:9dd7c64b4a64 9 *
jngarlitos 1:9dd7c64b4a64 10 * Description: Sum of the squares of the elements of a Q15 vector.
jngarlitos 1:9dd7c64b4a64 11 *
jngarlitos 1:9dd7c64b4a64 12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
jngarlitos 1:9dd7c64b4a64 13 *
jngarlitos 1:9dd7c64b4a64 14 * Redistribution and use in source and binary forms, with or without
jngarlitos 1:9dd7c64b4a64 15 * modification, are permitted provided that the following conditions
jngarlitos 1:9dd7c64b4a64 16 * are met:
jngarlitos 1:9dd7c64b4a64 17 * - Redistributions of source code must retain the above copyright
jngarlitos 1:9dd7c64b4a64 18 * notice, this list of conditions and the following disclaimer.
jngarlitos 1:9dd7c64b4a64 19 * - Redistributions in binary form must reproduce the above copyright
jngarlitos 1:9dd7c64b4a64 20 * notice, this list of conditions and the following disclaimer in
jngarlitos 1:9dd7c64b4a64 21 * the documentation and/or other materials provided with the
jngarlitos 1:9dd7c64b4a64 22 * distribution.
jngarlitos 1:9dd7c64b4a64 23 * - Neither the name of ARM LIMITED nor the names of its contributors
jngarlitos 1:9dd7c64b4a64 24 * may be used to endorse or promote products derived from this
jngarlitos 1:9dd7c64b4a64 25 * software without specific prior written permission.
jngarlitos 1:9dd7c64b4a64 26 *
jngarlitos 1:9dd7c64b4a64 27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
jngarlitos 1:9dd7c64b4a64 28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
jngarlitos 1:9dd7c64b4a64 29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
jngarlitos 1:9dd7c64b4a64 30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
jngarlitos 1:9dd7c64b4a64 31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
jngarlitos 1:9dd7c64b4a64 32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
jngarlitos 1:9dd7c64b4a64 33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
jngarlitos 1:9dd7c64b4a64 34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
jngarlitos 1:9dd7c64b4a64 35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
jngarlitos 1:9dd7c64b4a64 36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
jngarlitos 1:9dd7c64b4a64 37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
jngarlitos 1:9dd7c64b4a64 38 * POSSIBILITY OF SUCH DAMAGE.
jngarlitos 1:9dd7c64b4a64 39 * -------------------------------------------------------------------- */
jngarlitos 1:9dd7c64b4a64 40
jngarlitos 1:9dd7c64b4a64 41 #include "arm_math.h"
jngarlitos 1:9dd7c64b4a64 42
jngarlitos 1:9dd7c64b4a64 43 /**
jngarlitos 1:9dd7c64b4a64 44 * @ingroup groupStats
jngarlitos 1:9dd7c64b4a64 45 */
jngarlitos 1:9dd7c64b4a64 46
jngarlitos 1:9dd7c64b4a64 47 /**
jngarlitos 1:9dd7c64b4a64 48 * @addtogroup power
jngarlitos 1:9dd7c64b4a64 49 * @{
jngarlitos 1:9dd7c64b4a64 50 */
jngarlitos 1:9dd7c64b4a64 51
jngarlitos 1:9dd7c64b4a64 52 /**
jngarlitos 1:9dd7c64b4a64 53 * @brief Sum of the squares of the elements of a Q15 vector.
jngarlitos 1:9dd7c64b4a64 54 * @param[in] *pSrc points to the input vector
jngarlitos 1:9dd7c64b4a64 55 * @param[in] blockSize length of the input vector
jngarlitos 1:9dd7c64b4a64 56 * @param[out] *pResult sum of the squares value returned here
jngarlitos 1:9dd7c64b4a64 57 * @return none.
jngarlitos 1:9dd7c64b4a64 58 *
jngarlitos 1:9dd7c64b4a64 59 * @details
jngarlitos 1:9dd7c64b4a64 60 * <b>Scaling and Overflow Behavior:</b>
jngarlitos 1:9dd7c64b4a64 61 *
jngarlitos 1:9dd7c64b4a64 62 * \par
jngarlitos 1:9dd7c64b4a64 63 * The function is implemented using a 64-bit internal accumulator.
jngarlitos 1:9dd7c64b4a64 64 * The input is represented in 1.15 format.
jngarlitos 1:9dd7c64b4a64 65 * Intermediate multiplication yields a 2.30 format, and this
jngarlitos 1:9dd7c64b4a64 66 * result is added without saturation to a 64-bit accumulator in 34.30 format.
jngarlitos 1:9dd7c64b4a64 67 * With 33 guard bits in the accumulator, there is no risk of overflow, and the
jngarlitos 1:9dd7c64b4a64 68 * full precision of the intermediate multiplication is preserved.
jngarlitos 1:9dd7c64b4a64 69 * Finally, the return result is in 34.30 format.
jngarlitos 1:9dd7c64b4a64 70 *
jngarlitos 1:9dd7c64b4a64 71 */
jngarlitos 1:9dd7c64b4a64 72
jngarlitos 1:9dd7c64b4a64 73 void arm_power_q15(
jngarlitos 1:9dd7c64b4a64 74 q15_t * pSrc,
jngarlitos 1:9dd7c64b4a64 75 uint32_t blockSize,
jngarlitos 1:9dd7c64b4a64 76 q63_t * pResult)
jngarlitos 1:9dd7c64b4a64 77 {
jngarlitos 1:9dd7c64b4a64 78 q63_t sum = 0; /* Temporary result storage */
jngarlitos 1:9dd7c64b4a64 79
jngarlitos 1:9dd7c64b4a64 80 #ifndef ARM_MATH_CM0_FAMILY
jngarlitos 1:9dd7c64b4a64 81
jngarlitos 1:9dd7c64b4a64 82 /* Run the below code for Cortex-M4 and Cortex-M3 */
jngarlitos 1:9dd7c64b4a64 83
jngarlitos 1:9dd7c64b4a64 84 q31_t in32; /* Temporary variable to store input value */
jngarlitos 1:9dd7c64b4a64 85 q15_t in16; /* Temporary variable to store input value */
jngarlitos 1:9dd7c64b4a64 86 uint32_t blkCnt; /* loop counter */
jngarlitos 1:9dd7c64b4a64 87
jngarlitos 1:9dd7c64b4a64 88
jngarlitos 1:9dd7c64b4a64 89 /* loop Unrolling */
jngarlitos 1:9dd7c64b4a64 90 blkCnt = blockSize >> 2u;
jngarlitos 1:9dd7c64b4a64 91
jngarlitos 1:9dd7c64b4a64 92 /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
jngarlitos 1:9dd7c64b4a64 93 ** a second loop below computes the remaining 1 to 3 samples. */
jngarlitos 1:9dd7c64b4a64 94 while(blkCnt > 0u)
jngarlitos 1:9dd7c64b4a64 95 {
jngarlitos 1:9dd7c64b4a64 96 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
jngarlitos 1:9dd7c64b4a64 97 /* Compute Power and then store the result in a temporary variable, sum. */
jngarlitos 1:9dd7c64b4a64 98 in32 = *__SIMD32(pSrc)++;
jngarlitos 1:9dd7c64b4a64 99 sum = __SMLALD(in32, in32, sum);
jngarlitos 1:9dd7c64b4a64 100 in32 = *__SIMD32(pSrc)++;
jngarlitos 1:9dd7c64b4a64 101 sum = __SMLALD(in32, in32, sum);
jngarlitos 1:9dd7c64b4a64 102
jngarlitos 1:9dd7c64b4a64 103 /* Decrement the loop counter */
jngarlitos 1:9dd7c64b4a64 104 blkCnt--;
jngarlitos 1:9dd7c64b4a64 105 }
jngarlitos 1:9dd7c64b4a64 106
jngarlitos 1:9dd7c64b4a64 107 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
jngarlitos 1:9dd7c64b4a64 108 ** No loop unrolling is used. */
jngarlitos 1:9dd7c64b4a64 109 blkCnt = blockSize % 0x4u;
jngarlitos 1:9dd7c64b4a64 110
jngarlitos 1:9dd7c64b4a64 111 while(blkCnt > 0u)
jngarlitos 1:9dd7c64b4a64 112 {
jngarlitos 1:9dd7c64b4a64 113 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
jngarlitos 1:9dd7c64b4a64 114 /* Compute Power and then store the result in a temporary variable, sum. */
jngarlitos 1:9dd7c64b4a64 115 in16 = *pSrc++;
jngarlitos 1:9dd7c64b4a64 116 sum = __SMLALD(in16, in16, sum);
jngarlitos 1:9dd7c64b4a64 117
jngarlitos 1:9dd7c64b4a64 118 /* Decrement the loop counter */
jngarlitos 1:9dd7c64b4a64 119 blkCnt--;
jngarlitos 1:9dd7c64b4a64 120 }
jngarlitos 1:9dd7c64b4a64 121
jngarlitos 1:9dd7c64b4a64 122 #else
jngarlitos 1:9dd7c64b4a64 123
jngarlitos 1:9dd7c64b4a64 124 /* Run the below code for Cortex-M0 */
jngarlitos 1:9dd7c64b4a64 125
jngarlitos 1:9dd7c64b4a64 126 q15_t in; /* Temporary variable to store input value */
jngarlitos 1:9dd7c64b4a64 127 uint32_t blkCnt; /* loop counter */
jngarlitos 1:9dd7c64b4a64 128
jngarlitos 1:9dd7c64b4a64 129
jngarlitos 1:9dd7c64b4a64 130 /* Loop over blockSize number of values */
jngarlitos 1:9dd7c64b4a64 131 blkCnt = blockSize;
jngarlitos 1:9dd7c64b4a64 132
jngarlitos 1:9dd7c64b4a64 133 while(blkCnt > 0u)
jngarlitos 1:9dd7c64b4a64 134 {
jngarlitos 1:9dd7c64b4a64 135 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
jngarlitos 1:9dd7c64b4a64 136 /* Compute Power and then store the result in a temporary variable, sum. */
jngarlitos 1:9dd7c64b4a64 137 in = *pSrc++;
jngarlitos 1:9dd7c64b4a64 138 sum += ((q31_t) in * in);
jngarlitos 1:9dd7c64b4a64 139
jngarlitos 1:9dd7c64b4a64 140 /* Decrement the loop counter */
jngarlitos 1:9dd7c64b4a64 141 blkCnt--;
jngarlitos 1:9dd7c64b4a64 142 }
jngarlitos 1:9dd7c64b4a64 143
jngarlitos 1:9dd7c64b4a64 144 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
jngarlitos 1:9dd7c64b4a64 145
jngarlitos 1:9dd7c64b4a64 146 /* Store the results in 34.30 format */
jngarlitos 1:9dd7c64b4a64 147 *pResult = sum;
jngarlitos 1:9dd7c64b4a64 148 }
jngarlitos 1:9dd7c64b4a64 149
jngarlitos 1:9dd7c64b4a64 150 /**
jngarlitos 1:9dd7c64b4a64 151 * @} end of power group
jngarlitos 1:9dd7c64b4a64 152 */