jsonpath scanner: reentrant scanner
[pgsql.git] / src / common / pg_prng.c
blob15b39411a9b692b92df897fe7e0a1e6c046fb696
1 /*-------------------------------------------------------------------------
3 * Pseudo-Random Number Generator
5 * We use Blackman and Vigna's xoroshiro128** 1.0 algorithm
6 * to have a small, fast PRNG suitable for generating reasonably
7 * good-quality 64-bit data. This should not be considered
8 * cryptographically strong, however.
10 * About these generators: https://prng.di.unimi.it/
11 * See also https://en.wikipedia.org/wiki/List_of_random_number_generators
13 * Copyright (c) 2021-2024, PostgreSQL Global Development Group
15 * src/common/pg_prng.c
17 *-------------------------------------------------------------------------
20 #include "c.h"
22 #include <math.h>
24 #include "common/pg_prng.h"
25 #include "port/pg_bitutils.h"
27 /* X/Open (XSI) requires <math.h> to provide M_PI, but core POSIX does not */
28 #ifndef M_PI
29 #define M_PI 3.14159265358979323846
30 #endif
33 /* process-wide state vector */
34 pg_prng_state pg_global_prng_state;
38 * 64-bit rotate left
40 static inline uint64
41 rotl(uint64 x, int bits)
43 return (x << bits) | (x >> (64 - bits));
47 * The basic xoroshiro128** algorithm.
48 * Generates and returns a 64-bit uniformly distributed number,
49 * updating the state vector for next time.
51 * Note: the state vector must not be all-zeroes, as that is a fixed point.
53 static uint64
54 xoroshiro128ss(pg_prng_state *state)
56 uint64 s0 = state->s0,
57 sx = state->s1 ^ s0,
58 val = rotl(s0 * 5, 7) * 9;
60 /* update state */
61 state->s0 = rotl(s0, 24) ^ sx ^ (sx << 16);
62 state->s1 = rotl(sx, 37);
64 return val;
68 * We use this generator just to fill the xoroshiro128** state vector
69 * from a 64-bit seed.
71 static uint64
72 splitmix64(uint64 *state)
74 /* state update */
75 uint64 val = (*state += UINT64CONST(0x9E3779B97f4A7C15));
77 /* value extraction */
78 val = (val ^ (val >> 30)) * UINT64CONST(0xBF58476D1CE4E5B9);
79 val = (val ^ (val >> 27)) * UINT64CONST(0x94D049BB133111EB);
81 return val ^ (val >> 31);
85 * Initialize the PRNG state from a 64-bit integer,
86 * taking care that we don't produce all-zeroes.
88 void
89 pg_prng_seed(pg_prng_state *state, uint64 seed)
91 state->s0 = splitmix64(&seed);
92 state->s1 = splitmix64(&seed);
93 /* Let's just make sure we didn't get all-zeroes */
94 (void) pg_prng_seed_check(state);
98 * Initialize the PRNG state from a double in the range [-1.0, 1.0],
99 * taking care that we don't produce all-zeroes.
101 void
102 pg_prng_fseed(pg_prng_state *state, double fseed)
104 /* Assume there's about 52 mantissa bits; the sign contributes too. */
105 int64 seed = ((double) ((UINT64CONST(1) << 52) - 1)) * fseed;
107 pg_prng_seed(state, (uint64) seed);
111 * Validate a PRNG seed value.
113 bool
114 pg_prng_seed_check(pg_prng_state *state)
117 * If the seeding mechanism chanced to produce all-zeroes, insert
118 * something nonzero. Anything would do; use Knuth's LCG parameters.
120 if (unlikely(state->s0 == 0 && state->s1 == 0))
122 state->s0 = UINT64CONST(0x5851F42D4C957F2D);
123 state->s1 = UINT64CONST(0x14057B7EF767814F);
126 /* As a convenience for the pg_prng_strong_seed macro, return true */
127 return true;
131 * Select a random uint64 uniformly from the range [0, PG_UINT64_MAX].
133 uint64
134 pg_prng_uint64(pg_prng_state *state)
136 return xoroshiro128ss(state);
140 * Select a random uint64 uniformly from the range [rmin, rmax].
141 * If the range is empty, rmin is always produced.
143 uint64
144 pg_prng_uint64_range(pg_prng_state *state, uint64 rmin, uint64 rmax)
146 uint64 val;
148 if (likely(rmax > rmin))
151 * Use bitmask rejection method to generate an offset in 0..range.
152 * Each generated val is less than twice "range", so on average we
153 * should not have to iterate more than twice.
155 uint64 range = rmax - rmin;
156 uint32 rshift = 63 - pg_leftmost_one_pos64(range);
160 val = xoroshiro128ss(state) >> rshift;
161 } while (val > range);
163 else
164 val = 0;
166 return rmin + val;
170 * Select a random int64 uniformly from the range [PG_INT64_MIN, PG_INT64_MAX].
172 int64
173 pg_prng_int64(pg_prng_state *state)
175 return (int64) xoroshiro128ss(state);
179 * Select a random int64 uniformly from the range [0, PG_INT64_MAX].
181 int64
182 pg_prng_int64p(pg_prng_state *state)
184 return (int64) (xoroshiro128ss(state) & UINT64CONST(0x7FFFFFFFFFFFFFFF));
188 * Select a random int64 uniformly from the range [rmin, rmax].
189 * If the range is empty, rmin is always produced.
191 int64
192 pg_prng_int64_range(pg_prng_state *state, int64 rmin, int64 rmax)
194 int64 val;
196 if (likely(rmax > rmin))
198 uint64 uval;
201 * Use pg_prng_uint64_range(). Can't simply pass it rmin and rmax,
202 * since (uint64) rmin will be larger than (uint64) rmax if rmin < 0.
204 uval = (uint64) rmin +
205 pg_prng_uint64_range(state, 0, (uint64) rmax - (uint64) rmin);
208 * Safely convert back to int64, avoiding implementation-defined
209 * behavior for values larger than PG_INT64_MAX. Modern compilers
210 * will reduce this to a simple assignment.
212 if (uval > PG_INT64_MAX)
213 val = (int64) (uval - PG_INT64_MIN) + PG_INT64_MIN;
214 else
215 val = (int64) uval;
217 else
218 val = rmin;
220 return val;
224 * Select a random uint32 uniformly from the range [0, PG_UINT32_MAX].
226 uint32
227 pg_prng_uint32(pg_prng_state *state)
230 * Although xoroshiro128** is not known to have any weaknesses in
231 * randomness of low-order bits, we prefer to use the upper bits of its
232 * result here and below.
234 uint64 v = xoroshiro128ss(state);
236 return (uint32) (v >> 32);
240 * Select a random int32 uniformly from the range [PG_INT32_MIN, PG_INT32_MAX].
242 int32
243 pg_prng_int32(pg_prng_state *state)
245 uint64 v = xoroshiro128ss(state);
247 return (int32) (v >> 32);
251 * Select a random int32 uniformly from the range [0, PG_INT32_MAX].
253 int32
254 pg_prng_int32p(pg_prng_state *state)
256 uint64 v = xoroshiro128ss(state);
258 return (int32) (v >> 33);
262 * Select a random double uniformly from the range [0.0, 1.0).
264 * Note: if you want a result in the range (0.0, 1.0], the standard way
265 * to get that is "1.0 - pg_prng_double(state)".
267 double
268 pg_prng_double(pg_prng_state *state)
270 uint64 v = xoroshiro128ss(state);
273 * As above, assume there's 52 mantissa bits in a double. This result
274 * could round to 1.0 if double's precision is less than that; but we
275 * assume IEEE float arithmetic elsewhere in Postgres, so this seems OK.
277 return ldexp((double) (v >> (64 - 52)), -52);
281 * Select a random double from the normal distribution with
282 * mean = 0.0 and stddev = 1.0.
284 * To get a result from a different normal distribution use
285 * STDDEV * pg_prng_double_normal() + MEAN
287 * Uses https://en.wikipedia.org/wiki/Box%E2%80%93Muller_transform
289 double
290 pg_prng_double_normal(pg_prng_state *state)
292 double u1,
297 * pg_prng_double generates [0, 1), but for the basic version of the
298 * Box-Muller transform the two uniformly distributed random numbers are
299 * expected to be in (0, 1]; in particular we'd better not compute log(0).
301 u1 = 1.0 - pg_prng_double(state);
302 u2 = 1.0 - pg_prng_double(state);
304 /* Apply Box-Muller transform to get one normal-valued output */
305 z0 = sqrt(-2.0 * log(u1)) * sin(2.0 * M_PI * u2);
306 return z0;
310 * Select a random boolean value.
312 bool
313 pg_prng_bool(pg_prng_state *state)
315 uint64 v = xoroshiro128ss(state);
317 return (bool) (v >> 63);