arm_mat_scale_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_mat_scale_f32.c    
00009 *    
00010 * Description:  Multiplies a floating-point matrix by a scalar.    
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 MatrixScale Matrix Scale        
00049  *        
00050  * Multiplies a matrix by a scalar.  This is accomplished by multiplying each element in the        
00051  * matrix by the scalar.  For example:        
00052  * \image html MatrixScale.gif "Matrix Scaling of a 3 x 3 matrix"        
00053  *        
00054  * The function checks to make sure that the input and output matrices are of the same size.        
00055  *        
00056  * In the fixed-point Q15 and Q31 functions, <code>scale</code> is represented by        
00057  * a fractional multiplication <code>scaleFract</code> and an arithmetic shift <code>shift</code>.        
00058  * The shift allows the gain of the scaling operation to exceed 1.0.        
00059  * The overall scale factor applied to the fixed-point data is        
00060  * <pre>        
00061  *     scale = scaleFract * 2^shift.        
00062  * </pre>        
00063  */
00064 
00065 /**        
00066  * @addtogroup MatrixScale        
00067  * @{        
00068  */
00069 
00070 /**        
00071  * @brief Floating-point matrix scaling.        
00072  * @param[in]       *pSrc points to input matrix structure        
00073  * @param[in]       scale scale factor to be applied         
00074  * @param[out]      *pDst points to output matrix structure        
00075  * @return          The function returns either <code>ARM_MATH_SIZE_MISMATCH</code>         
00076  * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.        
00077  *        
00078  */
00079 
00080 arm_status arm_mat_scale_f32(
00081   const arm_matrix_instance_f32 * pSrc,
00082   float32_t scale,
00083   arm_matrix_instance_f32 * pDst)
00084 {
00085   float32_t *pIn = pSrc->pData;                  /* input data matrix pointer */
00086   float32_t *pOut = pDst->pData;                 /* output data matrix pointer */
00087   uint32_t numSamples;                           /* total number of elements in the matrix */
00088   uint32_t blkCnt;                               /* loop counters */
00089   arm_status status;                             /* status of matrix scaling     */
00090 
00091 #ifndef ARM_MATH_CM0_FAMILY
00092 
00093   float32_t in1, in2, in3, in4;                  /* temporary variables */
00094   float32_t out1, out2, out3, out4;              /* temporary variables */
00095 
00096 #endif //      #ifndef ARM_MATH_CM0_FAMILY
00097 
00098 #ifdef ARM_MATH_MATRIX_CHECK
00099   /* Check for matrix mismatch condition */
00100   if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols))
00101   {
00102     /* Set status as ARM_MATH_SIZE_MISMATCH */
00103     status = ARM_MATH_SIZE_MISMATCH;
00104   }
00105   else
00106 #endif /*    #ifdef ARM_MATH_MATRIX_CHECK    */
00107   {
00108     /* Total number of samples in the input matrix */
00109     numSamples = (uint32_t) pSrc->numRows * pSrc->numCols;
00110 
00111 #ifndef ARM_MATH_CM0_FAMILY
00112 
00113     /* Run the below code for Cortex-M4 and Cortex-M3 */
00114 
00115     /* Loop Unrolling */
00116     blkCnt = numSamples >> 2;
00117 
00118     /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00119      ** a second loop below computes the remaining 1 to 3 samples. */
00120     while(blkCnt > 0u)
00121     {
00122       /* C(m,n) = A(m,n) * scale */
00123       /* Scaling and results are stored in the destination buffer. */
00124       in1 = pIn[0];
00125       in2 = pIn[1];
00126       in3 = pIn[2];
00127       in4 = pIn[3];
00128 
00129       out1 = in1 * scale;
00130       out2 = in2 * scale;
00131       out3 = in3 * scale;
00132       out4 = in4 * scale;
00133 
00134 
00135       pOut[0] = out1;
00136       pOut[1] = out2;
00137       pOut[2] = out3;
00138       pOut[3] = out4;
00139 
00140       /* update pointers to process next sampels */
00141       pIn += 4u;
00142       pOut += 4u;
00143 
00144       /* Decrement the numSamples loop counter */
00145       blkCnt--;
00146     }
00147 
00148     /* If the numSamples is not a multiple of 4, compute any remaining output samples here.    
00149      ** No loop unrolling is used. */
00150     blkCnt = numSamples % 0x4u;
00151 
00152 #else
00153 
00154     /* Run the below code for Cortex-M0 */
00155 
00156     /* Initialize blkCnt with number of samples */
00157     blkCnt = numSamples;
00158 
00159 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00160 
00161     while(blkCnt > 0u)
00162     {
00163       /* C(m,n) = A(m,n) * scale */
00164       /* The results are stored in the destination buffer. */
00165       *pOut++ = (*pIn++) * scale;
00166 
00167       /* Decrement the loop counter */
00168       blkCnt--;
00169     }
00170 
00171     /* Set status as ARM_MATH_SUCCESS */
00172     status = ARM_MATH_SUCCESS;
00173   }
00174 
00175   /* Return to application */
00176   return (status);
00177 }
00178 
00179 /**        
00180  * @} end of MatrixScale group        
00181  */