arm_abs_q7.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_abs_q7.c    
00009 *    
00010 * Description:  Q7 vector absolute value.    
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  * @addtogroup BasicAbs        
00049  * @{        
00050  */
00051 
00052 /**        
00053  * @brief Q7 vector absolute value.        
00054  * @param[in]       *pSrc points to the input buffer        
00055  * @param[out]      *pDst points to the output buffer        
00056  * @param[in]       blockSize number of samples in each vector        
00057  * @return none.        
00058  *    
00059  * \par Conditions for optimum performance    
00060  *  Input and output buffers should be aligned by 32-bit    
00061  *    
00062  *        
00063  * <b>Scaling and Overflow Behavior:</b>        
00064  * \par        
00065  * The function uses saturating arithmetic.        
00066  * The Q7 value -1 (0x80) will be saturated to the maximum allowable positive value 0x7F.        
00067  */
00068 
00069 void arm_abs_q7(
00070   q7_t * pSrc,
00071   q7_t * pDst,
00072   uint32_t blockSize)
00073 {
00074   uint32_t blkCnt;                               /* loop counter */
00075   q7_t in;                                       /* Input value1 */
00076 
00077 #ifndef ARM_MATH_CM0_FAMILY
00078 
00079   /* Run the below code for Cortex-M4 and Cortex-M3 */
00080   q31_t in1, in2, in3, in4;                      /* temporary input variables */
00081   q31_t out1, out2, out3, out4;                  /* temporary output variables */
00082 
00083   /*loop Unrolling */
00084   blkCnt = blockSize >> 2u;
00085 
00086   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00087    ** a second loop below computes the remaining 1 to 3 samples. */
00088   while(blkCnt > 0u)
00089   {
00090     /* C = |A| */
00091     /* Read inputs */
00092     in1 = (q31_t) * pSrc;
00093     in2 = (q31_t) * (pSrc + 1);
00094     in3 = (q31_t) * (pSrc + 2);
00095 
00096     /* find absolute value */
00097     out1 = (in1 > 0) ? in1 : (q31_t)__QSUB8(0, in1);
00098 
00099     /* read input */
00100     in4 = (q31_t) * (pSrc + 3);
00101 
00102     /* find absolute value */
00103     out2 = (in2 > 0) ? in2 : (q31_t)__QSUB8(0, in2);
00104 
00105     /* store result to destination */
00106     *pDst = (q7_t) out1;
00107 
00108     /* find absolute value */
00109     out3 = (in3 > 0) ? in3 : (q31_t)__QSUB8(0, in3);
00110 
00111     /* find absolute value */
00112     out4 = (in4 > 0) ? in4 : (q31_t)__QSUB8(0, in4);
00113 
00114     /* store result to destination */
00115     *(pDst + 1) = (q7_t) out2;
00116 
00117     /* store result to destination */
00118     *(pDst + 2) = (q7_t) out3;
00119 
00120     /* store result to destination */
00121     *(pDst + 3) = (q7_t) out4;
00122 
00123     /* update pointers to process next samples */
00124     pSrc += 4u;
00125     pDst += 4u;
00126 
00127     /* Decrement the loop counter */
00128     blkCnt--;
00129   }
00130 
00131   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.    
00132    ** No loop unrolling is used. */
00133   blkCnt = blockSize % 0x4u;
00134 #else
00135 
00136   /* Run the below code for Cortex-M0 */
00137   blkCnt = blockSize;
00138 
00139 #endif //      #define ARM_MATH_CM0_FAMILY
00140 
00141   while(blkCnt > 0u)
00142   {
00143     /* C = |A| */
00144     /* Read the input */
00145     in = *pSrc++;
00146 
00147     /* Store the Absolute result in the destination buffer */
00148     *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? 0x7f : -in);
00149 
00150     /* Decrement the loop counter */
00151     blkCnt--;
00152   }
00153 }
00154 
00155 /**        
00156  * @} end of BasicAbs group        
00157  */