📄 mantissa.c
字号:
/* * The olsr.org Optimized Link-State Routing daemon(olsrd) * Copyright (c) 2004, Andreas T鴑nesen(andreto@olsr.org) * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * Neither the name of olsr.org, olsrd nor the names of its * contributors may be used to endorse or promote products derived * from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. * * Visit http://www.olsr.org for more information. * * If you find this software useful feel free to make a donation * to the project. For more information see the website or contact * the copyright holders. * * $Id: mantissa.c,v 1.9 2007/08/28 20:45:17 bernd67 Exp $ */#include "mantissa.h"/** *Function that converts a double to a mantissa/exponent *product as described in RFC3626: * * value = C*(1+a/16)*2^b [in seconds] * * where a is the integer represented by the four highest bits of the * field and b the integer represented by the four lowest bits of the * field. * *@param interval the time interval to process * *@return a 8-bit mantissa/exponent product */olsr_u8_tdouble_to_me(const double interval){ olsr_u8_t a, b; /* It is sufficent to compare the integer part since we test on >=. * So we have now only a floating point division and the rest of the loop * are only integer operations. */ const unsigned int unscaled_interval = interval / VTIME_SCALE_FACTOR; b = 0; while (unscaled_interval >= (1U << b)) { b++; } if(b == 0) { a = 1; b = 0; } else { b--; if (b > 15) { a = 15; b = 15; } else { /* And again some maths simplification from the former version: * a = 16 * ((double)interval / (VTIME_SCALE_FACTOR * (double)(1 << b)) - 1) * Since interval is already double: * a = 16 * (interval / (VTIME_SCALE_FACTOR * (double)(1 << b)) - 1) * first, we can get rid of parentheses and change the * to a / * a = 16 * (interval / VTIME_SCALE_FACTOR / (double)(1 << b) - 1) * then we make an integer addition from the floating point addition * a = (int)(16.0 * interval / VTIME_SCALE_FACTOR / (double)(1 << b)) - 16 * and we loose an unnecessary cast * a = (int)(16.0 * interval / VTIME_SCALE_FACTOR / (1 << b)) - 16 */ a = (int)(16.0 * interval / VTIME_SCALE_FACTOR / (1 << b)) - 16; b += a >> 4; a &= 0x0f; } } //printf("Generated mantissa/exponent(%d/%d): %d from %f\n", a, b, (olsr_u8_t) (a*16+b), interval); //printf("Resolves back to: %f\n", me_to_double(((a<<4)|(b&0x0F)))); return (a << 4) | (b & 0x0F);}/** * Function for converting a mantissa/exponent 8bit value back * to double as described in RFC3626: * * value = C*(1+a/16)*2^b [in seconds] * * where a is the integer represented by the four highest bits of the * field and b the integer represented by the four lowest bits of the * field. * * me is the 8 bit mantissa/exponent value * * To avoid expensive floating maths, we transform the equation: * value = C * (1 + a / 16) * 2^b * first, we make an int addition from the floating point addition: * value = C * ((16 + a) / 16) * 2^b * then we get rid of a pair of parentheses * value = C * (16 + a) / 16 * 2^b * and now we make an int multiplication from the floating point one * value = C * (16 + a) * 2^b / 16 * so that we can make a shift from the multiplication * value = C * ((16 + a) << b) / 16 * and sionce C and 16 are constants * value = ((16 + a) << b) * C / 16 */double me_to_double(const olsr_u8_t me){ const olsr_u8_t a = me >> 4; const olsr_u8_t b = me & 0x0F; return ((16 + a) << b) * VTIME_SCALE_FACTOR / 16.0;}
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -