arm_cmplx_mag_squared_q31.c

00001 /* ----------------------------------------------------------------------
00002  * Project:      CMSIS DSP Library
00003  * Title:        arm_cmplx_mag_squared_q31.c
00004  * Description:  Q31 complex magnitude squared
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_squared
00037  * @{
00038  */
00039 
00040 
00041 /**
00042  * @brief  Q31 complex magnitude squared
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 1.31 by 1.31 multiplications and finally output is converted into 3.29 format.
00051  * Input down scaling is not required.
00052  */
00053 
00054 void arm_cmplx_mag_squared_q31(
00055   q31_t * pSrc,
00056   q31_t * pDst,
00057   uint32_t numSamples)
00058 {
00059   q31_t real, imag;                              /* Temporary variables to store real and imaginary values */
00060   q31_t acc0, acc1;                              /* Accumulators */
00061 
00062 #if defined (ARM_MATH_DSP)
00063 
00064   /* Run the below code for Cortex-M4 and Cortex-M3 */
00065   uint32_t blkCnt;                               /* loop counter */
00066 
00067   /* loop Unrolling */
00068   blkCnt = numSamples >> 2U;
00069 
00070   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
00071    ** a second loop below computes the remaining 1 to 3 samples. */
00072   while (blkCnt > 0U)
00073   {
00074     /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
00075     real = *pSrc++;
00076     imag = *pSrc++;
00077     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00078     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00079     /* store the result in 3.29 format in the destination buffer. */
00080     *pDst++ = acc0 + acc1;
00081 
00082     real = *pSrc++;
00083     imag = *pSrc++;
00084     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00085     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00086     /* store the result in 3.29 format in the destination buffer. */
00087     *pDst++ = acc0 + acc1;
00088 
00089     real = *pSrc++;
00090     imag = *pSrc++;
00091     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00092     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00093     /* store the result in 3.29 format in the destination buffer. */
00094     *pDst++ = acc0 + acc1;
00095 
00096     real = *pSrc++;
00097     imag = *pSrc++;
00098     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00099     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00100     /* store the result in 3.29 format in the destination buffer. */
00101     *pDst++ = acc0 + acc1;
00102 
00103     /* Decrement the loop counter */
00104     blkCnt--;
00105   }
00106 
00107   /* If the numSamples is not a multiple of 4, compute any remaining output samples here.
00108    ** No loop unrolling is used. */
00109   blkCnt = numSamples % 0x4U;
00110 
00111   while (blkCnt > 0U)
00112   {
00113     /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
00114     real = *pSrc++;
00115     imag = *pSrc++;
00116     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00117     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00118     /* store the result in 3.29 format in the destination buffer. */
00119     *pDst++ = acc0 + acc1;
00120 
00121     /* Decrement the loop counter */
00122     blkCnt--;
00123   }
00124 
00125 #else
00126 
00127   /* Run the below code for Cortex-M0 */
00128 
00129   while (numSamples > 0U)
00130   {
00131     /* out = ((real * real) + (imag * imag)) */
00132     real = *pSrc++;
00133     imag = *pSrc++;
00134     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00135     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00136     /* store the result in 3.29 format in the destination buffer. */
00137     *pDst++ = acc0 + acc1;
00138 
00139     /* Decrement the loop counter */
00140     numSamples--;
00141   }
00142 
00143 #endif /* #if defined (ARM_MATH_DSP) */
00144 
00145 }
00146 
00147 /**
00148  * @} end of cmplx_mag_squared group
00149  */
00150