arm_mat_add_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_mat_add_f32.c    
00009 *    
00010 * Description:  Floating-point matrix addition    
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 groupMatrix        
00045  */
00046 
00047 /**        
00048  * @defgroup MatrixAdd Matrix Addition        
00049  *        
00050  * Adds two matrices.        
00051  * \image html MatrixAddition.gif "Addition of two 3 x 3 matrices"        
00052  *        
00053  * The functions check to make sure that        
00054  * <code>pSrcA</code>, <code>pSrcB</code>, and <code>pDst</code> have the same        
00055  * number of rows and columns.        
00056  */
00057 
00058 /**        
00059  * @addtogroup MatrixAdd        
00060  * @{        
00061  */
00062 
00063 
00064 /**        
00065  * @brief Floating-point matrix addition.        
00066  * @param[in]       *pSrcA points to the first input matrix structure        
00067  * @param[in]       *pSrcB points to the second input matrix structure        
00068  * @param[out]      *pDst points to output matrix structure        
00069  * @return          The function returns either        
00070  * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.        
00071  */
00072 
00073 arm_status arm_mat_add_f32(
00074   const arm_matrix_instance_f32 * pSrcA,
00075   const arm_matrix_instance_f32 * pSrcB,
00076   arm_matrix_instance_f32 * pDst)
00077 {
00078   float32_t *pIn1 = pSrcA->pData;                /* input data matrix pointer A  */
00079   float32_t *pIn2 = pSrcB->pData;                /* input data matrix pointer B  */
00080   float32_t *pOut = pDst->pData;                 /* output data matrix pointer   */
00081 
00082 #ifndef ARM_MATH_CM0_FAMILY
00083 
00084   float32_t inA1, inA2, inB1, inB2, out1, out2;  /* temporary variables */
00085 
00086 #endif //      #ifndef ARM_MATH_CM0_FAMILY
00087 
00088   uint32_t numSamples;                           /* total number of elements in the matrix  */
00089   uint32_t blkCnt;                               /* loop counters */
00090   arm_status status;                             /* status of matrix addition */
00091 
00092 #ifdef ARM_MATH_MATRIX_CHECK
00093   /* Check for matrix mismatch condition */
00094   if((pSrcA->numRows != pSrcB->numRows) ||
00095      (pSrcA->numCols != pSrcB->numCols) ||
00096      (pSrcA->numRows != pDst->numRows) || (pSrcA->numCols != pDst->numCols))
00097   {
00098     /* Set status as ARM_MATH_SIZE_MISMATCH */
00099     status = ARM_MATH_SIZE_MISMATCH;
00100   }
00101   else
00102 #endif
00103   {
00104 
00105     /* Total number of samples in the input matrix */
00106     numSamples = (uint32_t) pSrcA->numRows * pSrcA->numCols;
00107 
00108 #ifndef ARM_MATH_CM0_FAMILY
00109 
00110     /* Loop unrolling */
00111     blkCnt = numSamples >> 2u;
00112 
00113     /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00114      ** a second loop below computes the remaining 1 to 3 samples. */
00115     while(blkCnt > 0u)
00116     {
00117       /* C(m,n) = A(m,n) + B(m,n) */
00118       /* Add and then store the results in the destination buffer. */
00119       /* Read values from source A */
00120       inA1 = pIn1[0];
00121 
00122       /* Read values from source B */
00123       inB1 = pIn2[0];
00124 
00125       /* Read values from source A */
00126       inA2 = pIn1[1];
00127 
00128       /* out = sourceA + sourceB */
00129       out1 = inA1 + inB1;
00130 
00131       /* Read values from source B */
00132       inB2 = pIn2[1];
00133 
00134       /* Read values from source A */
00135       inA1 = pIn1[2];
00136 
00137       /* out = sourceA + sourceB */
00138       out2 = inA2 + inB2;
00139 
00140       /* Read values from source B */
00141       inB1 = pIn2[2];
00142 
00143       /* Store result in destination */
00144       pOut[0] = out1;
00145       pOut[1] = out2;
00146 
00147       /* Read values from source A */
00148       inA2 = pIn1[3];
00149 
00150       /* Read values from source B */
00151       inB2 = pIn2[3];
00152 
00153       /* out = sourceA + sourceB */
00154       out1 = inA1 + inB1;
00155 
00156       /* out = sourceA + sourceB */
00157       out2 = inA2 + inB2;
00158 
00159       /* Store result in destination */
00160       pOut[2] = out1;
00161 
00162       /* Store result in destination */
00163       pOut[3] = out2;
00164 
00165 
00166       /* update pointers to process next sampels */
00167       pIn1 += 4u;
00168       pIn2 += 4u;
00169       pOut += 4u;
00170       /* Decrement the loop counter */
00171       blkCnt--;
00172     }
00173 
00174     /* If the numSamples is not a multiple of 4, compute any remaining output samples here.    
00175      ** No loop unrolling is used. */
00176     blkCnt = numSamples % 0x4u;
00177 
00178 #else
00179 
00180     /* Run the below code for Cortex-M0 */
00181 
00182     /* Initialize blkCnt with number of samples */
00183     blkCnt = numSamples;
00184 
00185 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00186 
00187     while(blkCnt > 0u)
00188     {
00189       /* C(m,n) = A(m,n) + B(m,n) */
00190       /* Add and then store the results in the destination buffer. */
00191       *pOut++ = (*pIn1++) + (*pIn2++);
00192 
00193       /* Decrement the loop counter */
00194       blkCnt--;
00195     }
00196 
00197     /* set status as ARM_MATH_SUCCESS */
00198     status = ARM_MATH_SUCCESS;
00199 
00200   }
00201 
00202   /* Return to application */
00203   return (status);
00204 }
00205 
00206 /**        
00207  * @} end of MatrixAdd group        
00208  */