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_dot_prod_q15.c 00009 * 00010 * Description: Q15 dot product. 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 * Version 0.0.7 2010/06/10 00030 * Misra-C changes done 00031 * -------------------------------------------------------------------- */ 00032 00033 #include "arm_math.h" 00034 00061 void arm_dot_prod_q15( 00062 q15_t * pSrcA, 00063 q15_t * pSrcB, 00064 uint32_t blockSize, 00065 q63_t * result) 00066 { 00067 q63_t sum = 0; /* Temporary result storage */ 00068 uint32_t blkCnt; /* loop counter */ 00069 00070 #ifndef ARM_MATH_CM0 00071 00072 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00073 00074 00075 /*loop Unrolling */ 00076 blkCnt = blockSize >> 2u; 00077 00078 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00079 ** a second loop below computes the remaining 1 to 3 samples. */ 00080 while(blkCnt > 0u) 00081 { 00082 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00083 /* Calculate dot product and then store the result in a temporary buffer. */ 00084 sum = __SMLALD(*__SIMD32(pSrcA)++, *__SIMD32(pSrcB)++, sum); 00085 sum = __SMLALD(*__SIMD32(pSrcA)++, *__SIMD32(pSrcB)++, sum); 00086 00087 /* Decrement the loop counter */ 00088 blkCnt--; 00089 } 00090 00091 /* If the blockSize is not a multiple of 4, compute any remaining output samples here. 00092 ** No loop unrolling is used. */ 00093 blkCnt = blockSize % 0x4u; 00094 00095 while(blkCnt > 0u) 00096 { 00097 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00098 /* Calculate dot product and then store the results in a temporary buffer. */ 00099 sum = __SMLALD(*pSrcA++, *pSrcB++, sum); 00100 00101 /* Decrement the loop counter */ 00102 blkCnt--; 00103 } 00104 00105 00106 #else 00107 00108 /* Run the below code for Cortex-M0 */ 00109 00110 /* Initialize blkCnt with number of samples */ 00111 blkCnt = blockSize; 00112 00113 while(blkCnt > 0u) 00114 { 00115 /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */ 00116 /* Calculate dot product and then store the results in a temporary buffer. */ 00117 sum += (q63_t) ((q31_t) * pSrcA++ * *pSrcB++); 00118 00119 /* Decrement the loop counter */ 00120 blkCnt--; 00121 } 00122 00123 #endif /* #ifndef ARM_MATH_CM0 */ 00124 00125 /* Store the result in the destination buffer in 34.30 format */ 00126 *result = sum; 00127 00128 } 00129