arm_add_q31.c

00001 /* ----------------------------------------------------------------------
00002  * Project:      CMSIS DSP Library
00003  * Title:        arm_add_q31.c
00004  * Description:  Q31 vector addition
00005  *
00006  * $Date:        27. January 2017
00007  * $Revision:    V.1.5.1
00008  *
00009  * Target Processor: Cortex-M cores
00010  * -------------------------------------------------------------------- */
00011 /*
00012  * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved.
00013  *
00014  * SPDX-License-Identifier: Apache-2.0
00015  *
00016  * Licensed under the Apache License, Version 2.0 (the License); you may
00017  * not use this file except in compliance with the License.
00018  * You may obtain a copy of the License at
00019  *
00020  * www.apache.org/licenses/LICENSE-2.0
00021  *
00022  * Unless required by applicable law or agreed to in writing, software
00023  * distributed under the License is distributed on an AS IS BASIS, WITHOUT
00024  * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
00025  * See the License for the specific language governing permissions and
00026  * limitations under the License.
00027  */
00028 
00029 #include "arm_math.h"
00030 
00031 /**
00032  * @ingroup groupMath
00033  */
00034 
00035 /**
00036  * @addtogroup BasicAdd
00037  * @{
00038  */
00039 
00040 
00041 /**
00042  * @brief Q31 vector addition.
00043  * @param[in]       *pSrcA points to the first input vector
00044  * @param[in]       *pSrcB points to the second input vector
00045  * @param[out]      *pDst points to the output vector
00046  * @param[in]       blockSize number of samples in each vector
00047  * @return none.
00048  *
00049  * <b>Scaling and Overflow Behavior:</b>
00050  * \par
00051  * The function uses saturating arithmetic.
00052  * Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] will be saturated.
00053  */
00054 
00055 void arm_add_q31(
00056   q31_t * pSrcA,
00057   q31_t * pSrcB,
00058   q31_t * pDst,
00059   uint32_t blockSize)
00060 {
00061   uint32_t blkCnt;                               /* loop counter */
00062 
00063 #if defined (ARM_MATH_DSP)
00064 
00065 /* Run the below code for Cortex-M4 and Cortex-M3 */
00066   q31_t inA1, inA2, inA3, inA4;
00067   q31_t inB1, inB2, inB3, inB4;
00068 
00069   /*loop Unrolling */
00070   blkCnt = blockSize >> 2U;
00071 
00072   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
00073    ** a second loop below computes the remaining 1 to 3 samples. */
00074   while (blkCnt > 0U)
00075   {
00076     /* C = A + B */
00077     /* Add and then store the results in the destination buffer. */
00078     inA1 = *pSrcA++;
00079     inA2 = *pSrcA++;
00080     inB1 = *pSrcB++;
00081     inB2 = *pSrcB++;
00082 
00083     inA3 = *pSrcA++;
00084     inA4 = *pSrcA++;
00085     inB3 = *pSrcB++;
00086     inB4 = *pSrcB++;
00087 
00088     *pDst++ = __QADD(inA1, inB1);
00089     *pDst++ = __QADD(inA2, inB2);
00090     *pDst++ = __QADD(inA3, inB3);
00091     *pDst++ = __QADD(inA4, inB4);
00092 
00093     /* Decrement the loop counter */
00094     blkCnt--;
00095   }
00096 
00097   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
00098    ** No loop unrolling is used. */
00099   blkCnt = blockSize % 0x4U;
00100 
00101   while (blkCnt > 0U)
00102   {
00103     /* C = A + B */
00104     /* Add and then store the results in the destination buffer. */
00105     *pDst++ = __QADD(*pSrcA++, *pSrcB++);
00106 
00107     /* Decrement the loop counter */
00108     blkCnt--;
00109   }
00110 
00111 #else
00112 
00113   /* Run the below code for Cortex-M0 */
00114 
00115 
00116 
00117   /* Initialize blkCnt with number of samples */
00118   blkCnt = blockSize;
00119 
00120   while (blkCnt > 0U)
00121   {
00122     /* C = A + B */
00123     /* Add and then store the results in the destination buffer. */
00124     *pDst++ = (q31_t) clip_q63_to_q31((q63_t) * pSrcA++ + *pSrcB++);
00125 
00126     /* Decrement the loop counter */
00127     blkCnt--;
00128   }
00129 
00130 #endif /* #if defined (ARM_MATH_DSP) */
00131 
00132 }
00133 
00134 /**
00135  * @} end of BasicAdd group
00136  */
00137