CMSIS DSP Library from CMSIS 2.0. See http://www.onarm.com/cmsis/ for full details
Dependents: K22F_DSP_Matrix_least_square BNO055-ELEC3810 1BNO055 ECE4180Project--Slave2 ... more
arm_dot_prod_q7.c
00001 /* ---------------------------------------------------------------------- 00002 * Copyright (C) 2010 ARM Limited. All rights reserved. 00003 * 00004 * $Date: 29. November 2010 00005 * $Revision: V1.0.3 00006 * 00007 * Project: CMSIS DSP Library 00008 * Title: arm_dot_prod_q7.c 00009 * 00010 * Description: Q7 dot product. 00011 * 00012 * Target Processor: Cortex-M4/Cortex-M3 00013 * 00014 * Version 1.0.3 2010/11/29 00015 * Re-organized the CMSIS folders and updated documentation. 00016 * 00017 * Version 1.0.2 2010/11/11 00018 * Documentation updated. 00019 * 00020 * Version 1.0.1 2010/10/05 00021 * Production release and review comments incorporated. 00022 * 00023 * Version 1.0.0 2010/09/20 00024 * Production release and review comments incorporated. 00025 * 00026 * Version 0.0.7 2010/06/10 00027 * Misra-C changes done 00028 * -------------------------------------------------------------------- */ 00029 00030 #include "arm_math.h" 00031 00032 /** 00033 * @ingroup groupMath 00034 */ 00035 00036 /** 00037 * @addtogroup dot_prod 00038 * @{ 00039 */ 00040 00041 /** 00042 * @brief Dot product of Q7 vectors. 00043 * @param[in] *pSrcA points to the first input vector 00044 * @param[in] *pSrcB points to the second input vector 00045 * @param[in] blockSize number of samples in each vector 00046 * @param[out] *result output result returned here 00047 * @return none. 00048 * 00049 * <b>Scaling and Overflow Behavior:</b> 00050 * \par 00051 * The intermediate multiplications are in 1.7 x 1.7 = 2.14 format and these 00052 * results are added to an accumulator in 18.14 format. 00053 * Nonsaturating additions are used and there is no danger of wrap around as long as 00054 * the vectors are less than 2^18 elements long. 00055 * The return result is in 18.14 format. 00056 */ 00057 00058 void arm_dot_prod_q7( 00059 q7_t * pSrcA, 00060 q7_t * pSrcB, 00061 uint32_t blockSize, 00062 q31_t * result) 00063 { 00064 q31_t input1, input2; /* Temporary variables to store input */ 00065 q15_t in1, in2; /* Temporary variables to store input */ 00066 q31_t sum = 0; /* Temporary variables to store output */ 00067 uint32_t blkCnt; /* loop counter */ 00068 00069 00070 00071 /*loop Unrolling */ 00072 blkCnt = blockSize >> 2u; 00073 00074 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00075 ** a second loop below computes the remaining 1 to 3 samples. */ 00076 while(blkCnt > 0u) 00077 { 00078 /* Reading two inputs of SrcA buffer and packing */ 00079 in1 = (q15_t) * pSrcA++; 00080 in2 = (q15_t) * pSrcA++; 00081 input1 = ((q31_t) in1 & 0x0000FFFF) | ((q31_t) in2 << 16); 00082 00083 /* Reading two inputs of SrcB buffer and packing */ 00084 in1 = (q15_t) * pSrcB++; 00085 in2 = (q15_t) * pSrcB++; 00086 input2 = ((q31_t) in1 & 0x0000FFFF) | ((q31_t) in2 << 16); 00087 00088 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00089 /* Perform Dot product of 2 packed inputs using SMLALD and store the result in a temporary variable. */ 00090 sum = __SMLAD(input1, input2, sum); 00091 00092 /* Reading two inputs of SrcA buffer and packing */ 00093 in1 = (q15_t) * pSrcA++; 00094 in2 = (q15_t) * pSrcA++; 00095 input1 = ((q31_t) in1 & 0x0000FFFF) | ((q31_t) in2 << 16); 00096 00097 /* Reading two inputs of SrcB buffer and packing */ 00098 in1 = (q15_t) * pSrcB++; 00099 in2 = (q15_t) * pSrcB++; 00100 input2 = ((q31_t) in1 & 0x0000FFFF) | ((q31_t) in2 << 16); 00101 00102 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00103 /* Perform Dot product of 2 packed inputs using SMLALD and store the result in a temporary variable. */ 00104 sum = __SMLAD(input1, input2, sum); 00105 00106 00107 00108 /* Decrement the loop counter */ 00109 blkCnt--; 00110 } 00111 00112 /* If the blockSize is not a multiple of 4, compute any remaining output samples here. 00113 ** No loop unrolling is used. */ 00114 blkCnt = blockSize % 0x4u; 00115 00116 while(blkCnt > 0u) 00117 { 00118 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00119 /* Dot product and then store the results in a temporary buffer. */ 00120 sum = __SMLAD(*pSrcA++, *pSrcB++, sum); 00121 00122 /* Decrement the loop counter */ 00123 blkCnt--; 00124 } 00125 00126 /* Store the result in the destination buffer in 18.14 format */ 00127 *result = sum; 00128 } 00129 00130 /** 00131 * @} end of dot_prod group 00132 */
Generated on Tue Jul 12 2022 14:13:52 by 1.7.2