include/gmx_random.h

   1 /*
   2  *
   3  *                This source code is part of
   4  *
   5  *                 G   R   O   M   A   C   S
   6  *
   7  *          GROningen MAchine for Chemical Simulations
   8  *
   9  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  10  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  11  * Copyright (c) 2001-2008, The GROMACS development team,
  12  * check out http://www.gromacs.org for more information.
  13
  14  * This program is free software; you can redistribute it and/or
  15  * modify it under the terms of the GNU General Public License
  16  * as published by the Free Software Foundation; either version 2
  17  * of the License, or (at your option) any later version.
  18  *
  19  * If you want to redistribute modifications, please consider that
  20  * scientific software is very special. Version control is crucial -
  21  * bugs must be traceable. We will be happy to consider code for
  22  * inclusion in the official distribution, but derived work must not
  23  * be called official GROMACS. Details are found in the README & COPYING
  24  * files - if they are missing, get the official version at www.gromacs.org.
  25  *
  26  * To help us fund GROMACS development, we humbly ask that you cite
  27  * the papers on the package - you can find them in the top README file.
  28  *
  29  * For more info, check our website at http://www.gromacs.org
  30  *
  31  * And Hey:
  32  * Gallium Rubidium Oxygen Manganese Argon Carbon Silicon
  33  */
  34
  35 #ifndef _GMX_RANDOM_H_
  36 #define _GMX_RANDOM_H_
  37
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42
  43 #include <stdio.h>
  44 #include <types/simple.h>
  45
  46 /*! \brief Abstract datatype for a random number generator
  47  *
  48  * This is a handle to the full state of a random number generator.
  49  * You can not access anything inside the gmx_rng structure outside this
  50  * file.
  51  */
  52 typedef struct gmx_rng *
  53 gmx_rng_t;
  54
  55
  56 /*! \brief Returns the size of the RNG integer data structure
  57  *
  58  * Returns the size of the RNG integer data structure.
  59  * \threadsafe Yes.
  60  */
  61 int
  62 gmx_rng_n(void);
  63
  64
  65 /*! \brief Create a new RNG, seeded from a single integer.
  66  *
  67  * If you dont want to pick a seed, just call it as
  68  * rng=gmx_rng_init(gmx_rng_make_seed()) to seed it from
  69  * the system time or a random device.
  70  *
  71  * \param seed Random seed, unsigned 32-bit integer.
  72  *
  73  * \return Reference to a random number generator, or NULL if there was an
  74  *         error.
  75  *
  76  * \threadsafe Yes.
  77  */
  78 gmx_rng_t
  79 gmx_rng_init(unsigned int seed);
  80
  81
  82 /*! \brief Generate a 'random' RNG seed.
  83  *
  84  * This routine tries to get a seed from /dev/random if present,
  85  * and if not it uses time-of-day and process id to generate one.
  86  *
  87  * \return 32-bit unsigned integer random seed.
  88  *
  89  * Tip: If you use this in your code, it is a good idea to write the
  90  * returned random seed to a logfile, so you can recreate the exact sequence
  91  * of random number if you need to reproduce your run later for one reason
  92  * or another.
  93  *
  94  * \threadsafe Yes.
  95  */
  96 unsigned int
  97 gmx_rng_make_seed(void);
  98
  99
 100 /*! \brief Initialize a RNG with 624 integers (>32 bits of entropy).
 101  *
 102  *  The Mersenne twister RNG used in Gromacs has an extremely long period,
 103  *  but when you only initialize it with a 32-bit integer there are only
 104  *  2^32 different possible sequences of number - much less than the generator
 105  *  is capable of.
 106  *
 107  *  If you really need the full entropy, this routine makes it possible to
 108  *  initialize the RNG with up to 624 32-bit integers, which will give you
 109  *  up to 2^19968 bits of entropy.
 110  *
 111  *  \param seed Array of unsigned integers to form a seed
 112  *  \param seed_length Number of integers in the array, up to 624 are used.
 113  *
 114  * \return Reference to a random number generator, or NULL if there was an
 115  *         error.
 116  *
 117  * \threadsafe Yes.
 118  */
 119 gmx_rng_t
 120 gmx_rng_init_array(unsigned int    seed[],
 121                    int             seed_length);
 122
 123
 124 /*! \brief Release resources of a RNG
 125  *
 126  *  This routine destroys a random number generator and releases all
 127  *  resources allocated by it.
 128  *
 129  *  \param rng Handle to random number generator previously returned by
 130  *                     gmx_rng_init() or gmx_rng_init_array().
 131  *
 132  * \threadsafe Function itself is threadsafe, but you should only destroy a
 133  *             certain RNG once (i.e. from one thread).
 134  */
 135 void
 136 gmx_rng_destroy(gmx_rng_t rng);
 137
 138
 139 /*! \brief Get the state of a RNG
 140  *
 141  *  This routine stores the random state in mt and mti, mt should have
 142  *  a size of at least 624, mt of 1.
 143  *
 144  *  \param rng Handle to random number generator previously returned by
 145  *                     gmx_rng_init() or gmx_rng_init_array().
 146  */
 147 void
 148 gmx_rng_get_state(gmx_rng_t rng, unsigned int *mt,int *mti);
 149
 150
 151 /*! \brief Set the state of a RNG
 152  *
 153  *  This routine sets the random state from mt and mti, mt should have
 154  *  a size of at least 624.
 155  *
 156  *  \param rng Handle to random number generator previously returned by
 157  *                     gmx_rng_init() or gmx_rng_init_array().
 158  */
 159 void
 160 gmx_rng_set_state(gmx_rng_t rng, unsigned int *mt,int mti);
 161
 162
 163 /*! \brief Random 32-bit integer from a uniform distribution
 164  *
 165  *  This routine returns a random integer from the random number generator
 166  *  provided, and updates the state of that RNG.
 167  *
 168  *  \param rng Handle to random number generator previously returned by
 169  *                     gmx_rng_init() or gmx_rng_init_array().
 170  *
 171  *  \return 32-bit unsigned integer from a uniform distribution.
 172  *
 173  *  \threadsafe Function yes, input data no. You should not call this function
 174  *              from two different threads using the same RNG handle at the
 175  *              same time. For performance reasons we cannot lock the handle
 176  *              with a mutex every time we need a random number - that would
 177  *              slow the routine down a factor 2-5. There are two simple
 178  *              solutions: either use a mutex and lock it before calling
 179  *              the function, or use a separate RNG handle for each thread.
 180  */
 181 unsigned int
 182 gmx_rng_uniform_uint32(gmx_rng_t rng);
 183
 184
 185 /*! \brief Random gmx_real_t 0<=x<1 from a uniform distribution
 186  *
 187  *  This routine returns a random floating-point number from the
 188  *  random number generator provided, and updates the state of that RNG.
 189  *
 190  *  \param rng Handle to random number generator previously returned by
 191  *                     gmx_rng_init() or gmx_rng_init_array().
 192  *
 193  *  \return floating-point number 0<=x<1 from a uniform distribution.
 194  *
 195  *  \threadsafe Function yes, input data no. You should not call this function
 196  *              from two different threads using the same RNG handle at the
 197  *              same time. For performance reasons we cannot lock the handle
 198  *              with a mutex every time we need a random number - that would
 199  *              slow the routine down a factor 2-5. There are two simple
 200  *              solutions: either use a mutex and lock it before calling
 201  *              the function, or use a separate RNG handle for each thread.
 202  */
 203 real
 204 gmx_rng_uniform_real(gmx_rng_t rng);
 205
 206
 207 /*! \brief Random gmx_real_t from a gaussian distribution
 208  *
 209  *  This routine returns a random floating-point number from the
 210  *  random number generator provided, and updates the state of that RNG.
 211  *
 212  *  The Box-Muller algorithm is used to provide gaussian random numbers. This
 213  *  is not the fastest known algorithm for gaussian numbers, but in contrast
 214  *  to the alternatives it is very well studied and you can trust the returned
 215  *  random numbers to have good properties and no correlations.
 216  *
 217  *  \param rng Handle to random number generator previously returned by
 218  *                        gmx_rng_init() or gmx_rng_init_array().
 219  *
 220  *  \return Gaussian random floating-point number with average 0.0 and
 221  *          standard deviation 1.0. You can get any average/mean you want
 222  *          by first multiplying with the desired average and then adding
 223  *          the average you want.
 224  *
 225  *  \threadsafe Function yes, input data no. You should not call this function
 226  *              from two different threads using the same RNG handle at the
 227  *              same time. For performance reasons we cannot lock the handle
 228  *              with a mutex every time we need a random number - that would
 229  *              slow the routine down a factor 2-5. There are two simple
 230  *              solutions: either use a mutex and lock it before calling
 231  *              the function, or use a separate RNG handle for each thread.
 232  *
 233  *  It works perfectly to mix calls to get uniform and gaussian random numbers
 234  *  from the same generator, but since it will affect the sequence of returned
 235  *  numbers it is probably better to use separate random number generator
 236  *  structures.
 237  */
 238 real
 239 gmx_rng_gaussian_real(gmx_rng_t rng);
 240
 241
 242
 243 /* Return a new gaussian random number with expectation value
 244  * 0.0 and standard deviation 1.0. This routine uses a table
 245  * lookup for maximum speed.
 246  *
 247  * WARNING: The lookup table is 16k by default, which means
 248  *          the granularity of the random numbers is coarser
 249  *          than what you get from gmx_rng_gauss_real().
 250  *          In most cases this is no problem whatsoever,
 251  *          and it is particularly true for BD/SD integration.
 252  *          Note that you will NEVER get any really extreme
 253  *          numbers: the maximum absolute value returned is
 254  *          4.0255485.
 255  *
 256  * threadsafe: yes
 257  */
 258 real
 259 gmx_rng_gaussian_table(gmx_rng_t rng);
 260
 261
 262 #endif /* _GMX_RANDOM_H_ */
 263