arm_cmplx_mag_q15.c

00001 /* ----------------------------------------------------------------------
00002  * Project:      CMSIS DSP Library
00003  * Title:        arm_cmplx_mag_q15.c
00004  * Description:  Q15 complex magnitude
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 groupCmplxMath
00033  */
00034 
00035 /**
00036  * @addtogroup cmplx_mag
00037  * @{
00038  */
00039 
00040 
00041 /**
00042  * @brief  Q15 complex magnitude
00043  * @param  *pSrc points to the complex input vector
00044  * @param  *pDst points to the real output vector
00045  * @param  numSamples number of complex samples in the input vector
00046  * @return none.
00047  *
00048  * <b>Scaling and Overflow Behavior:</b>
00049  * \par
00050  * The function implements 1.15 by 1.15 multiplications and finally output is converted into 2.14 format.
00051  */
00052 
00053 void arm_cmplx_mag_q15(
00054   q15_t * pSrc,
00055   q15_t * pDst,
00056   uint32_t numSamples)
00057 {
00058   q31_t acc0, acc1;                              /* Accumulators */
00059 
00060 #if defined (ARM_MATH_DSP)
00061 
00062   /* Run the below code for Cortex-M4 and Cortex-M3 */
00063   uint32_t blkCnt;                               /* loop counter */
00064   q31_t in1, in2, in3, in4;
00065   q31_t acc2, acc3;
00066 
00067 
00068   /*loop Unrolling */
00069   blkCnt = numSamples >> 2U;
00070 
00071   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
00072    ** a second loop below computes the remaining 1 to 3 samples. */
00073   while (blkCnt > 0U)
00074   {
00075 
00076     /* C[0] = sqrt(A[0] * A[0] + A[1] * A[1]) */
00077     in1 = *__SIMD32(pSrc)++;
00078     in2 = *__SIMD32(pSrc)++;
00079     in3 = *__SIMD32(pSrc)++;
00080     in4 = *__SIMD32(pSrc)++;
00081 
00082     acc0 = __SMUAD(in1, in1);
00083     acc1 = __SMUAD(in2, in2);
00084     acc2 = __SMUAD(in3, in3);
00085     acc3 = __SMUAD(in4, in4);
00086 
00087     /* store the result in 2.14 format in the destination buffer. */
00088     arm_sqrt_q15((q15_t) ((acc0) >> 17), pDst++);
00089     arm_sqrt_q15((q15_t) ((acc1) >> 17), pDst++);
00090     arm_sqrt_q15((q15_t) ((acc2) >> 17), pDst++);
00091     arm_sqrt_q15((q15_t) ((acc3) >> 17), pDst++);
00092 
00093     /* Decrement the loop counter */
00094     blkCnt--;
00095   }
00096 
00097   /* If the numSamples is not a multiple of 4, compute any remaining output samples here.
00098    ** No loop unrolling is used. */
00099   blkCnt = numSamples % 0x4U;
00100 
00101   while (blkCnt > 0U)
00102   {
00103     /* C[0] = sqrt(A[0] * A[0] + A[1] * A[1]) */
00104     in1 = *__SIMD32(pSrc)++;
00105     acc0 = __SMUAD(in1, in1);
00106 
00107     /* store the result in 2.14 format in the destination buffer. */
00108     arm_sqrt_q15((q15_t) (acc0 >> 17), pDst++);
00109 
00110     /* Decrement the loop counter */
00111     blkCnt--;
00112   }
00113 
00114 #else
00115 
00116   /* Run the below code for Cortex-M0 */
00117   q15_t real, imag;                              /* Temporary variables to hold input values */
00118 
00119   while (numSamples > 0U)
00120   {
00121     /* out = sqrt(real * real + imag * imag) */
00122     real = *pSrc++;
00123     imag = *pSrc++;
00124 
00125     acc0 = (real * real);
00126     acc1 = (imag * imag);
00127 
00128     /* store the result in 2.14 format in the destination buffer. */
00129     arm_sqrt_q15((q15_t) (((q63_t) acc0 + acc1) >> 17), pDst++);
00130 
00131     /* Decrement the loop counter */
00132     numSamples--;
00133   }
00134 
00135 #endif /* #if defined (ARM_MATH_DSP) */
00136 
00137 }
00138 
00139 /**
00140  * @} end of cmplx_mag group
00141  */
00142