arm_dot_prod_f32.c

00001 /* ----------------------------------------------------------------------    
00002 * Copyright (C) 2010-2013 ARM Limited. All rights reserved.    
00003 *    
00004 * $Date:        17. January 2013
00005 * $Revision:    V1.4.1
00006 *    
00007 * Project:      CMSIS DSP Library    
00008 * Title:        arm_dot_prod_f32.c    
00009 *    
00010 * Description:  Floating-point dot product.    
00011 *    
00012 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
00013 *  
00014 * Redistribution and use in source and binary forms, with or without 
00015 * modification, are permitted provided that the following conditions
00016 * are met:
00017 *   - Redistributions of source code must retain the above copyright
00018 *     notice, this list of conditions and the following disclaimer.
00019 *   - Redistributions in binary form must reproduce the above copyright
00020 *     notice, this list of conditions and the following disclaimer in
00021 *     the documentation and/or other materials provided with the 
00022 *     distribution.
00023 *   - Neither the name of ARM LIMITED nor the names of its contributors
00024 *     may be used to endorse or promote products derived from this
00025 *     software without specific prior written permission.
00026 *
00027 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
00028 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
00029 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
00030 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 
00031 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
00032 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
00033 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
00034 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
00035 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
00036 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
00037 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
00038 * POSSIBILITY OF SUCH DAMAGE.  
00039 * ---------------------------------------------------------------------------- */
00040 
00041 #include "arm_math.h"
00042 
00043 /**    
00044  * @ingroup groupMath
00045  */
00046 
00047 /**
00048  * @defgroup dot_prod Vector Dot Product
00049  *
00050  * Computes the dot product of two vectors.
00051  * The vectors are multiplied element-by-element and then summed.
00052  *
00053  * <pre>
00054  *     sum = pSrcA[0]*pSrcB[0] + pSrcA[1]*pSrcB[1] + ... + pSrcA[blockSize-1]*pSrcB[blockSize-1]
00055  * </pre>     
00056  *
00057  * There are separate functions for floating-point, Q7, Q15, and Q31 data types.    
00058  */
00059 
00060 /**    
00061  * @addtogroup dot_prod    
00062  * @{    
00063  */
00064 
00065 /**    
00066  * @brief Dot product of floating-point vectors.    
00067  * @param[in]       *pSrcA points to the first input vector    
00068  * @param[in]       *pSrcB points to the second input vector    
00069  * @param[in]       blockSize number of samples in each vector    
00070  * @param[out]      *result output result returned here    
00071  * @return none.    
00072  */
00073 
00074 
00075 void arm_dot_prod_f32(
00076   float32_t * pSrcA,
00077   float32_t * pSrcB,
00078   uint32_t blockSize,
00079   float32_t * result)
00080 {
00081   float32_t sum = 0.0f;                          /* Temporary result storage */
00082   uint32_t blkCnt;                               /* loop counter */
00083 
00084 
00085 #ifndef ARM_MATH_CM0_FAMILY
00086 
00087 /* Run the below code for Cortex-M4 and Cortex-M3 */
00088   /*loop Unrolling */
00089   blkCnt = blockSize >> 2u;
00090 
00091   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00092    ** a second loop below computes the remaining 1 to 3 samples. */
00093   while(blkCnt > 0u)
00094   {
00095     /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
00096     /* Calculate dot product and then store the result in a temporary buffer */
00097     sum += (*pSrcA++) * (*pSrcB++);
00098     sum += (*pSrcA++) * (*pSrcB++);
00099     sum += (*pSrcA++) * (*pSrcB++);
00100     sum += (*pSrcA++) * (*pSrcB++);
00101 
00102     /* Decrement the loop counter */
00103     blkCnt--;
00104   }
00105 
00106   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.    
00107    ** No loop unrolling is used. */
00108   blkCnt = blockSize % 0x4u;
00109 
00110 #else
00111 
00112   /* Run the below code for Cortex-M0 */
00113 
00114   /* Initialize blkCnt with number of samples */
00115   blkCnt = blockSize;
00116 
00117 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00118 
00119 
00120   while(blkCnt > 0u)
00121   {
00122     /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
00123     /* Calculate dot product and then store the result in a temporary buffer. */
00124     sum += (*pSrcA++) * (*pSrcB++);
00125 
00126     /* Decrement the loop counter */
00127     blkCnt--;
00128   }
00129   /* Store the result back in the destination buffer */
00130   *result = sum;
00131 }
00132 
00133 /**    
00134  * @} end of dot_prod group    
00135  */