arm_add_q31.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_add_q31.c    
00009 *    
00010 * Description:  Q31 vector 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 groupMath    
00045  */
00046 
00047 /**    
00048  * @addtogroup BasicAdd    
00049  * @{    
00050  */
00051 
00052 
00053 /**    
00054  * @brief Q31 vector addition.    
00055  * @param[in]       *pSrcA points to the first input vector    
00056  * @param[in]       *pSrcB points to the second input vector    
00057  * @param[out]      *pDst points to the output vector    
00058  * @param[in]       blockSize number of samples in each vector    
00059  * @return none.    
00060  *    
00061  * <b>Scaling and Overflow Behavior:</b>    
00062  * \par    
00063  * The function uses saturating arithmetic.    
00064  * Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] will be saturated.    
00065  */
00066 
00067 void arm_add_q31(
00068   q31_t * pSrcA,
00069   q31_t * pSrcB,
00070   q31_t * pDst,
00071   uint32_t blockSize)
00072 {
00073   uint32_t blkCnt;                               /* loop counter */
00074 
00075 #ifndef ARM_MATH_CM0_FAMILY
00076 
00077 /* Run the below code for Cortex-M4 and Cortex-M3 */
00078   q31_t inA1, inA2, inA3, inA4;
00079   q31_t inB1, inB2, inB3, inB4;
00080 
00081   /*loop Unrolling */
00082   blkCnt = blockSize >> 2u;
00083 
00084   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00085    ** a second loop below computes the remaining 1 to 3 samples. */
00086   while(blkCnt > 0u)
00087   {
00088     /* C = A + B */
00089     /* Add and then store the results in the destination buffer. */
00090     inA1 = *pSrcA++;
00091     inA2 = *pSrcA++;
00092     inB1 = *pSrcB++;
00093     inB2 = *pSrcB++;
00094 
00095     inA3 = *pSrcA++;
00096     inA4 = *pSrcA++;
00097     inB3 = *pSrcB++;
00098     inB4 = *pSrcB++;
00099 
00100     *pDst++ = __QADD(inA1, inB1);
00101     *pDst++ = __QADD(inA2, inB2);
00102     *pDst++ = __QADD(inA3, inB3);
00103     *pDst++ = __QADD(inA4, inB4);
00104 
00105     /* Decrement the loop counter */
00106     blkCnt--;
00107   }
00108 
00109   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.    
00110    ** No loop unrolling is used. */
00111   blkCnt = blockSize % 0x4u;
00112 
00113   while(blkCnt > 0u)
00114   {
00115     /* C = A + B */
00116     /* Add and then store the results in the destination buffer. */
00117     *pDst++ = __QADD(*pSrcA++, *pSrcB++);
00118 
00119     /* Decrement the loop counter */
00120     blkCnt--;
00121   }
00122 
00123 #else
00124 
00125   /* Run the below code for Cortex-M0 */
00126 
00127 
00128 
00129   /* Initialize blkCnt with number of samples */
00130   blkCnt = blockSize;
00131 
00132   while(blkCnt > 0u)
00133   {
00134     /* C = A + B */
00135     /* Add and then store the results in the destination buffer. */
00136     *pDst++ = (q31_t) clip_q63_to_q31((q63_t) * pSrcA++ + *pSrcB++);
00137 
00138     /* Decrement the loop counter */
00139     blkCnt--;
00140   }
00141 
00142 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00143 
00144 }
00145 
00146 /**    
00147  * @} end of BasicAdd group    
00148  */