arm_cmplx_mag_squared_q15.c

00001 /* ----------------------------------------------------------------------    
00002 * Copyright (C) 2010-2014 ARM Limited. All rights reserved.    
00003 *    
00004 * $Date:        19. March 2015
00005 * $Revision:    V.1.4.5
00006 *    
00007 * Project:      CMSIS DSP Library    
00008 * Title:        arm_cmplx_mag_squared_q15.c    
00009 *    
00010 * Description:  Q15 complex magnitude squared.    
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 groupCmplxMath    
00045  */
00046 
00047 /**    
00048  * @addtogroup cmplx_mag_squared    
00049  * @{    
00050  */
00051 
00052 /**    
00053  * @brief  Q15 complex magnitude squared    
00054  * @param  *pSrc points to the complex input vector    
00055  * @param  *pDst points to the real output vector    
00056  * @param  numSamples number of complex samples in the input vector    
00057  * @return none.    
00058  *    
00059  * <b>Scaling and Overflow Behavior:</b>    
00060  * \par    
00061  * The function implements 1.15 by 1.15 multiplications and finally output is converted into 3.13 format.    
00062  */
00063 
00064 void arm_cmplx_mag_squared_q15(
00065   q15_t * pSrc,
00066   q15_t * pDst,
00067   uint32_t numSamples)
00068 {
00069   q31_t acc0, acc1;                              /* Accumulators */
00070 
00071 #ifndef ARM_MATH_CM0_FAMILY
00072 
00073   /* Run the below code for Cortex-M4 and Cortex-M3 */
00074   uint32_t blkCnt;                               /* loop counter */
00075   q31_t in1, in2, in3, in4;
00076   q31_t acc2, acc3;
00077 
00078   /*loop Unrolling */
00079   blkCnt = numSamples >> 2u;
00080 
00081   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00082    ** a second loop below computes the remaining 1 to 3 samples. */
00083   while(blkCnt > 0u)
00084   {
00085     /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
00086     in1 = *__SIMD32(pSrc)++;
00087     in2 = *__SIMD32(pSrc)++;
00088     in3 = *__SIMD32(pSrc)++;
00089     in4 = *__SIMD32(pSrc)++;
00090 
00091     acc0 = __SMUAD(in1, in1);
00092     acc1 = __SMUAD(in2, in2);
00093     acc2 = __SMUAD(in3, in3);
00094     acc3 = __SMUAD(in4, in4);
00095 
00096     /* store the result in 3.13 format in the destination buffer. */
00097     *pDst++ = (q15_t) (acc0 >> 17);
00098     *pDst++ = (q15_t) (acc1 >> 17);
00099     *pDst++ = (q15_t) (acc2 >> 17);
00100     *pDst++ = (q15_t) (acc3 >> 17);
00101 
00102     /* Decrement the loop counter */
00103     blkCnt--;
00104   }
00105 
00106   /* If the numSamples is not a multiple of 4, compute any remaining output samples here.    
00107    ** No loop unrolling is used. */
00108   blkCnt = numSamples % 0x4u;
00109 
00110   while(blkCnt > 0u)
00111   {
00112     /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
00113     in1 = *__SIMD32(pSrc)++;
00114     acc0 = __SMUAD(in1, in1);
00115 
00116     /* store the result in 3.13 format in the destination buffer. */
00117     *pDst++ = (q15_t) (acc0 >> 17);
00118 
00119     /* Decrement the loop counter */
00120     blkCnt--;
00121   }
00122 
00123 #else
00124 
00125   /* Run the below code for Cortex-M0 */
00126   q15_t real, imag;                              /* Temporary variables to store real and imaginary values */
00127 
00128   while(numSamples > 0u)
00129   {
00130     /* out = ((real * real) + (imag * imag)) */
00131     real = *pSrc++;
00132     imag = *pSrc++;
00133     acc0 = (real * real);
00134     acc1 = (imag * imag);
00135     /* store the result in 3.13 format in the destination buffer. */
00136     *pDst++ = (q15_t) (((q63_t) acc0 + acc1) >> 17);
00137 
00138     /* Decrement the loop counter */
00139     numSamples--;
00140   }
00141 
00142 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00143 
00144 }
00145 
00146 /**    
00147  * @} end of cmplx_mag_squared group    
00148  */