00001 /* ---------------------------------------------------------------------- 00002 * Copyright (C) 2010 ARM Limited. All rights reserved. 00003 * 00004 * $Date: 15. July 2011 00005 * $Revision: V1.0.10 00006 * 00007 * Project: CMSIS DSP Library 00008 * Title: arm_power_q31.c 00009 * 00010 * Description: Sum of the squares of the elements of a Q31 vector. 00011 * 00012 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0 00013 * 00014 * Version 1.0.10 2011/7/15 00015 * Big Endian support added and Merged M0 and M3/M4 Source code. 00016 * 00017 * Version 1.0.3 2010/11/29 00018 * Re-organized the CMSIS folders and updated documentation. 00019 * 00020 * Version 1.0.2 2010/11/11 00021 * Documentation updated. 00022 * 00023 * Version 1.0.1 2010/10/05 00024 * Production release and review comments incorporated. 00025 * 00026 * Version 1.0.0 2010/09/20 00027 * Production release and review comments incorporated. 00028 * -------------------------------------------------------------------- */ 00029 00030 #include "arm_math.h" 00031 00063 void arm_power_q31( 00064 q31_t * pSrc, 00065 uint32_t blockSize, 00066 q63_t * pResult) 00067 { 00068 q63_t sum = 0; /* Temporary result storage */ 00069 q31_t in; 00070 uint32_t blkCnt; /* loop counter */ 00071 00072 00073 #ifndef ARM_MATH_CM0 00074 00075 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00076 00077 /*loop Unrolling */ 00078 blkCnt = blockSize >> 2u; 00079 00080 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00081 ** a second loop below computes the remaining 1 to 3 samples. */ 00082 while(blkCnt > 0u) 00083 { 00084 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ 00085 /* Compute Power then shift intermediate results by 14 bits to maintain 16.48 format and then store the result in a temporary variable sum, providing 15 guard bits. */ 00086 in = *pSrc++; 00087 sum += ((q63_t) in * in) >> 14u; 00088 00089 in = *pSrc++; 00090 sum += ((q63_t) in * in) >> 14u; 00091 00092 in = *pSrc++; 00093 sum += ((q63_t) in * in) >> 14u; 00094 00095 in = *pSrc++; 00096 sum += ((q63_t) in * in) >> 14u; 00097 00098 /* Decrement the loop counter */ 00099 blkCnt--; 00100 } 00101 00102 /* If the blockSize is not a multiple of 4, compute any remaining output samples here. 00103 ** No loop unrolling is used. */ 00104 blkCnt = blockSize % 0x4u; 00105 00106 #else 00107 00108 /* Run the below code for Cortex-M0 */ 00109 00110 /* Loop over blockSize number of values */ 00111 blkCnt = blockSize; 00112 00113 #endif /* #ifndef ARM_MATH_CM0 */ 00114 00115 while(blkCnt > 0u) 00116 { 00117 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */ 00118 /* Compute Power and then store the result in a temporary variable, sum. */ 00119 in = *pSrc++; 00120 sum += ((q63_t) in * in) >> 14u; 00121 00122 /* Decrement the loop counter */ 00123 blkCnt--; 00124 } 00125 00126 /* Store the results in 16.48 format */ 00127 *pResult = sum; 00128 } 00129