00001 /* ---------------------------------------------------------------------- 00002 * Copyright (C) 2010 ARM Limited. All rights reserved. 00003 * 00004 * $Date: 15. July 2011 00005 * $Revision: V1.0.10 00006 * 00007 * Project: CMSIS DSP Library 00008 * Title: arm_cmplx_dot_prod_f32.c 00009 * 00010 * Description: Floating-point complex dot product 00011 * 00012 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0 00013 * 00014 * Version 1.0.10 2011/7/15 00015 * Big Endian support added and Merged M0 and M3/M4 Source code. 00016 * 00017 * Version 1.0.3 2010/11/29 00018 * Re-organized the CMSIS folders and updated documentation. 00019 * 00020 * Version 1.0.2 2010/11/11 00021 * Documentation updated. 00022 * 00023 * Version 1.0.1 2010/10/05 00024 * Production release and review comments incorporated. 00025 * 00026 * Version 1.0.0 2010/09/20 00027 * Production release and review comments incorporated. 00028 * ---------------------------------------------------------------------------- */ 00029 00030 #include "arm_math.h" 00031 00077 void arm_cmplx_dot_prod_f32( 00078 float32_t * pSrcA, 00079 float32_t * pSrcB, 00080 uint32_t numSamples, 00081 float32_t * realResult, 00082 float32_t * imagResult) 00083 { 00084 float32_t real_sum = 0.0f, imag_sum = 0.0f; /* Temporary result storage */ 00085 00086 #ifndef ARM_MATH_CM0 00087 00088 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00089 uint32_t blkCnt; /* loop counter */ 00090 00091 /*loop Unrolling */ 00092 blkCnt = numSamples >> 2u; 00093 00094 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00095 ** a second loop below computes the remaining 1 to 3 samples. */ 00096 while(blkCnt > 0u) 00097 { 00098 /* CReal = A[0]* B[0] + A[2]* B[2] + A[4]* B[4] + .....+ A[numSamples-2]* B[numSamples-2] */ 00099 real_sum += (*pSrcA++) * (*pSrcB++); 00100 /* CImag = A[1]* B[1] + A[3]* B[3] + A[5]* B[5] + .....+ A[numSamples-1]* B[numSamples-1] */ 00101 imag_sum += (*pSrcA++) * (*pSrcB++); 00102 00103 real_sum += (*pSrcA++) * (*pSrcB++); 00104 imag_sum += (*pSrcA++) * (*pSrcB++); 00105 00106 real_sum += (*pSrcA++) * (*pSrcB++); 00107 imag_sum += (*pSrcA++) * (*pSrcB++); 00108 00109 real_sum += (*pSrcA++) * (*pSrcB++); 00110 imag_sum += (*pSrcA++) * (*pSrcB++); 00111 00112 /* Decrement the loop counter */ 00113 blkCnt--; 00114 } 00115 00116 /* If the numSamples is not a multiple of 4, compute any remaining output samples here. 00117 ** No loop unrolling is used. */ 00118 blkCnt = numSamples % 0x4u; 00119 00120 while(blkCnt > 0u) 00121 { 00122 /* CReal = A[0]* B[0] + A[2]* B[2] + A[4]* B[4] + .....+ A[numSamples-2]* B[numSamples-2] */ 00123 real_sum += (*pSrcA++) * (*pSrcB++); 00124 /* CImag = A[1]* B[1] + A[3]* B[3] + A[5]* B[5] + .....+ A[numSamples-1]* B[numSamples-1] */ 00125 imag_sum += (*pSrcA++) * (*pSrcB++); 00126 00127 00128 /* Decrement the loop counter */ 00129 blkCnt--; 00130 } 00131 00132 #else 00133 00134 /* Run the below code for Cortex-M0 */ 00135 00136 while(numSamples > 0u) 00137 { 00138 /* CReal = A[0]* B[0] + A[2]* B[2] + A[4]* B[4] + .....+ A[numSamples-2]* B[numSamples-2] */ 00139 real_sum += (*pSrcA++) * (*pSrcB++); 00140 /* CImag = A[1]* B[1] + A[3]* B[3] + A[5]* B[5] + .....+ A[numSamples-1]* B[numSamples-1] */ 00141 imag_sum += (*pSrcA++) * (*pSrcB++); 00142 00143 00144 /* Decrement the loop counter */ 00145 numSamples--; 00146 } 00147 00148 #endif /* #ifndef ARM_MATH_CM0 */ 00149 00150 /* Store the real and imaginary results in the destination buffers */ 00151 *realResult = real_sum; 00152 *imagResult = imag_sum; 00153 } 00154