arm_cmplx_mult_cmplx_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_mult_cmplx_q15.c    
00009 *    
00010 * Description:  Q15 complex-by-complex multiplication    
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 CmplxByCmplxMult    
00049  * @{    
00050  */
00051 
00052 /**    
00053  * @brief  Q15 complex-by-complex multiplication    
00054  * @param[in]  *pSrcA points to the first input vector    
00055  * @param[in]  *pSrcB points to the second input vector    
00056  * @param[out]  *pDst  points to the output vector    
00057  * @param[in]  numSamples number of complex samples in each vector    
00058  * @return none.    
00059  *    
00060  * <b>Scaling and Overflow Behavior:</b>    
00061  * \par    
00062  * The function implements 1.15 by 1.15 multiplications and finally output is converted into 3.13 format.    
00063  */
00064 
00065 void arm_cmplx_mult_cmplx_q15(
00066   q15_t * pSrcA,
00067   q15_t * pSrcB,
00068   q15_t * pDst,
00069   uint32_t numSamples)
00070 {
00071   q15_t a, b, c, d;                              /* Temporary variables to store real and imaginary values */
00072 
00073 #ifndef ARM_MATH_CM0_FAMILY
00074 
00075   /* Run the below code for Cortex-M4 and Cortex-M3 */
00076   uint32_t blkCnt;                               /* loop counters */
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[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
00086     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
00087     a = *pSrcA++;
00088     b = *pSrcA++;
00089     c = *pSrcB++;
00090     d = *pSrcB++;
00091 
00092     /* store the result in 3.13 format in the destination buffer. */
00093     *pDst++ =
00094       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00095     /* store the result in 3.13 format in the destination buffer. */
00096     *pDst++ =
00097       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00098 
00099     a = *pSrcA++;
00100     b = *pSrcA++;
00101     c = *pSrcB++;
00102     d = *pSrcB++;
00103 
00104     /* store the result in 3.13 format in the destination buffer. */
00105     *pDst++ =
00106       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00107     /* store the result in 3.13 format in the destination buffer. */
00108     *pDst++ =
00109       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00110 
00111     a = *pSrcA++;
00112     b = *pSrcA++;
00113     c = *pSrcB++;
00114     d = *pSrcB++;
00115 
00116     /* store the result in 3.13 format in the destination buffer. */
00117     *pDst++ =
00118       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00119     /* store the result in 3.13 format in the destination buffer. */
00120     *pDst++ =
00121       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00122 
00123     a = *pSrcA++;
00124     b = *pSrcA++;
00125     c = *pSrcB++;
00126     d = *pSrcB++;
00127 
00128     /* store the result in 3.13 format in the destination buffer. */
00129     *pDst++ =
00130       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00131     /* store the result in 3.13 format in the destination buffer. */
00132     *pDst++ =
00133       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00134 
00135     /* Decrement the blockSize loop counter */
00136     blkCnt--;
00137   }
00138 
00139   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.    
00140    ** No loop unrolling is used. */
00141   blkCnt = numSamples % 0x4u;
00142 
00143   while(blkCnt > 0u)
00144   {
00145     /* C[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
00146     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
00147     a = *pSrcA++;
00148     b = *pSrcA++;
00149     c = *pSrcB++;
00150     d = *pSrcB++;
00151 
00152     /* store the result in 3.13 format in the destination buffer. */
00153     *pDst++ =
00154       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00155     /* store the result in 3.13 format in the destination buffer. */
00156     *pDst++ =
00157       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00158 
00159     /* Decrement the blockSize loop counter */
00160     blkCnt--;
00161   }
00162 
00163 #else
00164 
00165   /* Run the below code for Cortex-M0 */
00166 
00167   while(numSamples > 0u)
00168   {
00169     /* C[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
00170     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
00171     a = *pSrcA++;
00172     b = *pSrcA++;
00173     c = *pSrcB++;
00174     d = *pSrcB++;
00175 
00176     /* store the result in 3.13 format in the destination buffer. */
00177     *pDst++ =
00178       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
00179     /* store the result in 3.13 format in the destination buffer. */
00180     *pDst++ =
00181       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
00182 
00183     /* Decrement the blockSize loop counter */
00184     numSamples--;
00185   }
00186 
00187 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00188 
00189 }
00190 
00191 /**    
00192  * @} end of CmplxByCmplxMult group    
00193  */