arm_mult_f32.c

00001 /* ----------------------------------------------------------------------    
00002 * Copyright (C) 2010-2014 ARM Limited. All rights reserved.    
00003 *    
00004 * $Date:        19. March 2015
00005 * $Revision:    V.1.4.5
00006 *    
00007 * Project:      CMSIS DSP Library    
00008 * Title:        arm_mult_f32.c    
00009 *    
00010 * Description:  Floating-point vector multiplication.    
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 BasicMult Vector Multiplication        
00049  *        
00050  * Element-by-element multiplication of two vectors.        
00051  *        
00052  * <pre>        
00053  *     pDst[n] = pSrcA[n] * pSrcB[n],   0 <= n < blockSize.        
00054  * </pre>        
00055  *        
00056  * There are separate functions for floating-point, Q7, Q15, and Q31 data types.        
00057  */
00058 
00059 /**        
00060  * @addtogroup BasicMult        
00061  * @{        
00062  */
00063 
00064 /**        
00065  * @brief Floating-point vector multiplication.        
00066  * @param[in]       *pSrcA points to the first input vector        
00067  * @param[in]       *pSrcB points to the second input vector        
00068  * @param[out]      *pDst points to the output vector        
00069  * @param[in]       blockSize number of samples in each vector        
00070  * @return none.        
00071  */
00072 
00073 void arm_mult_f32(
00074   float32_t * pSrcA,
00075   float32_t * pSrcB,
00076   float32_t * pDst,
00077   uint32_t blockSize)
00078 {
00079   uint32_t blkCnt;                               /* loop counters */
00080 #ifndef ARM_MATH_CM0_FAMILY
00081 
00082   /* Run the below code for Cortex-M4 and Cortex-M3 */
00083   float32_t inA1, inA2, inA3, inA4;              /* temporary input variables */
00084   float32_t inB1, inB2, inB3, inB4;              /* temporary input variables */
00085   float32_t out1, out2, out3, out4;              /* temporary output variables */
00086 
00087   /* loop Unrolling */
00088   blkCnt = blockSize >> 2u;
00089 
00090   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.        
00091    ** a second loop below computes the remaining 1 to 3 samples. */
00092   while(blkCnt > 0u)
00093   {
00094     /* C = A * B */
00095     /* Multiply the inputs and store the results in output buffer */
00096     /* read sample from sourceA */
00097     inA1 = *pSrcA;
00098     /* read sample from sourceB */
00099     inB1 = *pSrcB;
00100     /* read sample from sourceA */
00101     inA2 = *(pSrcA + 1);
00102     /* read sample from sourceB */
00103     inB2 = *(pSrcB + 1);
00104 
00105     /* out = sourceA * sourceB */
00106     out1 = inA1 * inB1;
00107 
00108     /* read sample from sourceA */
00109     inA3 = *(pSrcA + 2);
00110     /* read sample from sourceB */
00111     inB3 = *(pSrcB + 2);
00112 
00113     /* out = sourceA * sourceB */
00114     out2 = inA2 * inB2;
00115 
00116     /* read sample from sourceA */
00117     inA4 = *(pSrcA + 3);
00118 
00119     /* store result to destination buffer */
00120     *pDst = out1;
00121 
00122     /* read sample from sourceB */
00123     inB4 = *(pSrcB + 3);
00124 
00125     /* out = sourceA * sourceB */
00126     out3 = inA3 * inB3;
00127 
00128     /* store result to destination buffer */
00129     *(pDst + 1) = out2;
00130 
00131     /* out = sourceA * sourceB */
00132     out4 = inA4 * inB4;
00133     /* store result to destination buffer */
00134     *(pDst + 2) = out3;
00135     /* store result to destination buffer */
00136     *(pDst + 3) = out4;
00137 
00138 
00139     /* update pointers to process next samples */
00140     pSrcA += 4u;
00141     pSrcB += 4u;
00142     pDst += 4u;
00143 
00144     /* Decrement the blockSize loop counter */
00145     blkCnt--;
00146   }
00147 
00148   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.        
00149    ** No loop unrolling is used. */
00150   blkCnt = blockSize % 0x4u;
00151 
00152 #else
00153 
00154   /* Run the below code for Cortex-M0 */
00155 
00156   /* Initialize blkCnt with number of samples */
00157   blkCnt = blockSize;
00158 
00159 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00160 
00161   while(blkCnt > 0u)
00162   {
00163     /* C = A * B */
00164     /* Multiply the inputs and store the results in output buffer */
00165     *pDst++ = (*pSrcA++) * (*pSrcB++);
00166 
00167     /* Decrement the blockSize loop counter */
00168     blkCnt--;
00169   }
00170 }
00171 
00172 /**        
00173  * @} end of BasicMult group        
00174  */