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_shift_q31.c 00009 * 00010 * Description: Shifts the elements of a Q31 vector by a specified number of bits. 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 * Version 0.0.7 2010/06/10 00030 * Misra-C changes done 00031 * -------------------------------------------------------------------- */ 00032 00033 #include "arm_math.h" 00034 00073 void arm_shift_q31( 00074 q31_t * pSrc, 00075 int8_t shiftBits, 00076 q31_t * pDst, 00077 uint32_t blockSize) 00078 { 00079 uint32_t blkCnt; /* loop counter */ 00080 uint8_t sign; /* Sign of shiftBits */ 00081 /* Getting the sign of shiftBits */ 00082 sign = (shiftBits & 0x80); 00083 00084 00085 #ifndef ARM_MATH_CM0 00086 00087 /* Run the below code for Cortex-M4 and Cortex-M3 */ 00088 /*loop Unrolling */ 00089 blkCnt = blockSize >> 2u; 00090 00091 00092 /* First part of the processing with loop unrolling. Compute 4 outputs at a time. 00093 ** a second loop below computes the remaining 1 to 3 samples. */ 00094 while(blkCnt > 0u) 00095 { 00096 /* C = A (>> or <<) shiftBits */ 00097 /* Shift the input and then store the results in the destination buffer. */ 00098 *pDst++ = (sign == 0u) ? clip_q63_to_q31((q63_t) * pSrc++ << shiftBits) : 00099 (*pSrc++ >> -shiftBits); 00100 *pDst++ = (sign == 0u) ? clip_q63_to_q31((q63_t) * pSrc++ << shiftBits) : 00101 (*pSrc++ >> -shiftBits); 00102 *pDst++ = (sign == 0u) ? clip_q63_to_q31((q63_t) * pSrc++ << shiftBits) : 00103 (*pSrc++ >> -shiftBits); 00104 *pDst++ = (sign == 0u) ? clip_q63_to_q31((q63_t) * pSrc++ << shiftBits) : 00105 (*pSrc++ >> -shiftBits); 00106 00107 /* Decrement the loop counter */ 00108 blkCnt--; 00109 } 00110 00111 /* If the blockSize is not a multiple of 4, compute any remaining output samples here. 00112 ** No loop unrolling is used. */ 00113 blkCnt = blockSize % 0x4u; 00114 00115 #else 00116 00117 /* Run the below code for Cortex-M0 */ 00118 00119 00120 /* Initialize blkCnt with number of samples */ 00121 blkCnt = blockSize; 00122 00123 #endif /* #ifndef ARM_MATH_CM0 */ 00124 00125 00126 while(blkCnt > 0u) 00127 { 00128 /* C = A (>> or <<) shiftBits */ 00129 /* Shift the input and then store the result in the destination buffer. */ 00130 *pDst++ = (sign == 0u) ? clip_q63_to_q31((q63_t) * pSrc++ << shiftBits) : 00131 (*pSrc++ >> -shiftBits); 00132 00133 /* Decrement the loop counter */ 00134 blkCnt--; 00135 } 00136 00137 } 00138