SpectMorph
|
#include <smpcg32rng.hh>
Public Member Functions | |
template<class SeedSeq > | |
Pcg32Rng (SeedSeq &seed_sequence) | |
Initialize and seed from seed_sequence. | |
Pcg32Rng (uint64_t offset, uint64_t sequence) | |
Initialize and seed by seeking to position offset within stream sequence. | |
Pcg32Rng () | |
Initialize and seed the generator from a system specific nondeterministic random source. | |
void | auto_seed () |
Seed the generator from a system specific nondeterministic random source. | |
void | seed (uint64_t offset, uint64_t sequence) |
Seed by seeking to position offset within stream sequence. | |
template<class SeedSeq > | |
void | seed (SeedSeq &seed_sequence) |
Seed the generator state from a seed_sequence. | |
uint32_t | random () |
Generate uniformly distributed 32 bit pseudo random number. | |
Pcg32Rng is a permutating linear congruential PRNG. At the core, this pseudo random number generator uses the well known linear congruential generator: 6364136223846793005 * accumulator + 1442695040888963407 mod 2^64. See also TAOCP by D. E. Knuth, section 3.3.4, table 1, line 26. For good statistical performance, the output function of the permuted congruential generator family is used as described on http://www.pcg-random.org/. Period length for this generator is 2^64, the specified seed offset chooses the position of the genrator and the seed sequence parameter can be used to choose from 2^63 distinct random sequences.