basic.h
/**
* \file basic.h
* \brief Portable types, math constants, and timing
* \author Pascal Getreuer <getreuer@cmla.ens-cachan.fr>
*
* This purpose of this file is to improve portability.
*
* Types \c uint8_t, \c uint16_t, \c uint32_t should be defined as
* unsigned integer types such that
* \li \c uint8_t is 8-bit, range 0 to 255
* \li \c uint16_t is 16-bit, range 0 to 65535
* \li \c uint32_t is 32-bit, range 0 to 4294967295
*
* Similarly, \c int8_t, \c int16_t, \c int32_t should be defined as
* signed integer types such that
* \li \c int8_t is 8-bit, range -128 to +127
* \li \c int16_t is 16-bit, range -32768 to +32767
* \li \c int32_t is 32-bit, range -2147483648 to +2147483647
*
* These definitions are implemented with types \c __int8, \c __int16,
* and \c __int32 under Windows and by including stdint.h under UNIX.
*
* To define the math constants, math.h is included, and any of the
* following that were not defined by math.h are defined here according
* to the values from Hart & Cheney.
* \li M_2PI = 2 pi = 6.28318530717958647692528676655900576
* \li M_PI = pi = 3.14159265358979323846264338327950288
* \li M_PI_2 = pi/2 = 1.57079632679489661923132169163975144
* \li M_PI_4 = pi/4 = 0.78539816339744830961566084581987572
* \li M_PI_8 = pi/8 = 0.39269908169872415480783042290993786
* \li M_SQRT2 = sqrt(2) = 1.41421356237309504880168872420969808
* \li M_1_SQRT2 = 1/sqrt(2) = 0.70710678118654752440084436210484904
* \li M_E = e = 2.71828182845904523536028747135266250
* \li M_LOG2E = log_2(e) = 1.44269504088896340735992468100189213
* \li M_LOG10E = log_10(e) = 0.43429448190325182765112891891660508
* \li M_LN2 = log_e(2) = 0.69314718055994530941723212145817657
* \li M_LN10 = log_e(10) = 2.30258509299404568401799145468436421
* \li M_EULER = Euler = 0.57721566490153286060651209008240243
* \li M_SQRT2PI = sqrt(2 pi) = 2.50662827463100050241576528481104525
*
* For precise timing, a function millisecond_timer() is defined.
*
*
* Copyright (c) 2010-2013, Pascal Getreuer
* All rights reserved.
*
* This program is free software: you can redistribute it and/or modify it
* under, at your option, the terms of the GNU General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version, or the terms of the
* simplified BSD license.
*
* You should have received a copy of these licenses along with this program.
* If not, see <http://www.gnu.org/licenses/> and
* <http://www.opensource.org/licenses/bsd-license.html>.
*/
#ifndef _BASIC_H_
#define _BASIC_H_
#include <math.h>
#include <stdlib.h>
/* Portable integer types */
#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
/* Windows system: Use __intN types to define uint8_t, etc. */
typedef unsigned __int8 uint8_t;
typedef unsigned __int16 uint16_t;
typedef unsigned __int32 uint32_t;
typedef __int8 int8_t;
typedef __int16 int16_t;
typedef __int32 int32_t;
#else
/* UNIX system: Use stdint to define uint8_t, etc. */
#include <stdint.h>
#endif
/* Math constants (Hart & Cheney) */
#ifndef M_2PI
/** \brief The constant 2 pi */
#define M_2PI 6.28318530717958647692528676655900576
#endif
#ifndef M_PI
/** \brief The constant pi */
#define M_PI 3.14159265358979323846264338327950288
#endif
#ifndef M_PI_2
/** \brief The constant pi/2 */
#define M_PI_2 1.57079632679489661923132169163975144
#endif
#ifndef M_PI_4
/** \brief The constant pi/4 */
#define M_PI_4 0.78539816339744830961566084581987572
#endif
#ifndef M_PI_8
/** \brief The constant pi/8 */
#define M_PI_8 0.39269908169872415480783042290993786
#endif
#ifndef M_SQRT2
/** \brief The constant sqrt(2) */
#define M_SQRT2 1.41421356237309504880168872420969808
#endif
#ifndef M_1_SQRT2
/** \brief The constant 1/sqrt(2) */
#define M_1_SQRT2 0.70710678118654752440084436210484904
#endif
#ifndef M_E
/** \brief The natural number */
#define M_E 2.71828182845904523536028747135266250
#endif
#ifndef M_LOG2E
/** \brief Log base 2 of the natural number */
#define M_LOG2E 1.44269504088896340735992468100189213
#endif
#ifndef M_LOG10E
/** \brief Log base 10 of the natural number */
#define M_LOG10E 0.43429448190325182765112891891660508
#endif
#ifndef M_LN2
/** \brief Natural log of 2 */
#define M_LN2 0.69314718055994530941723212145817657
#endif
#ifndef M_LN10
/** \brief Natural log of 10 */
#define M_LN10 2.30258509299404568401799145468436421
#endif
#ifndef M_EULER
/** \brief Euler number */
#define M_EULER 0.57721566490153286060651209008240243
#endif
#ifndef M_SQRT2PI
/** \brief The constant sqrt(2 pi) */
#define M_SQRT2PI 2.50662827463100050241576528481104525
#endif
/** \brief Round double X */
#define ROUND(X) (floor((X) + 0.5))
/** \brief Round float X */
#define ROUNDF(X) (floor((X) + 0.5f))
#ifdef __GNUC__
#ifndef ATTRIBUTE_UNUSED
/** \brief Macro for the unused attribue GNU extension */
#define ATTRIBUTE_UNUSED __attribute__((unused))
#endif
#ifndef ATTRIBUTE_ALWAYSINLINE
/** \brief Macro for the always inline attribue GNU extension */
#define ATTRIBUTE_ALWAYSINLINE __attribute__((always_inline))
#endif
#else
#define ATTRIBUTE_UNUSED
#define ATTRIBUTE_ALWAYSINLINE
#endif
/**
* \brief Millisecond-precision timer function
* \return Clock value in units of milliseconds
*
* This routine implements a timer with millisecond precision. In order to
* obtain timing at high resolution, platform-specific functions are needed:
*
* - On Windows systems, the GetSystemTime function is used.
* - On POSIX systems, the gettimeofday function is used.
*
* Otherwise as a fallback, time.h time is used, and in this case
* millisecond_timer() has only second accuracy. Preprocessor symbols are
* checked in attempt to detect whether the platform is POSIX or Windows and
* defines millisecond_timer() accordingly. A particular implementation can
* be forced by defining USE_GETSYSTEMTIME, USE_GETTIMEOFDAY, or USE_TIME.
*/
unsigned long millisecond_timer();
#endif /* _BASIC_H_ */
