XorwowRngData.hh File Reference

#include <cstdint>
#include "corecel/Assert.hh"
#include "corecel/Macros.hh"
#include "corecel/Types.hh"
#include "corecel/cont/Array.hh"
#include "corecel/cont/Span.hh"
#include "corecel/data/Collection.hh"

This graph shows which files directly or indirectly include this file:

Classes
struct	celeritas::XorwowRngParamsData< W, M >
	Persistent data for XORWOW generator. More...
struct	celeritas::XorwowRngInitializer
	Initialize an RNG. More...
struct	celeritas::XorwowState
	Individual RNG state. More...
struct	celeritas::XorwowRngStateData< W, M >
	XORWOW generator states for all threads. More...

Typedefs
using	celeritas::XorwowUInt = std::uint32_t
	32-bit unsigned integer type for xorwow
using	celeritas::XorwowSeed = Array< XorwowUInt, 1 >
	Seed type used to generate initial states for the RNG.

Functions
void	celeritas::initialize_xorwow (Span< XorwowState > state, XorwowSeed const &seed, StreamId stream)
	Initialize XORWOW states with well-distributed random data. More...
template<MemSpace M>
void	celeritas::resize (XorwowRngStateData< Ownership::value, M > *state, HostCRef< XorwowRngParamsData > const &params, StreamId stream, size_type size)
	Resize and seed the RNG states.

Function Documentation

initialize_xorwow()

void celeritas::initialize_xorwow	(	Span< XorwowState >	state,
		XorwowSeed const &	seed,
		StreamId	stream
	)

Initialize XORWOW states with well-distributed random data.

cuRAND and hipRAND implement functions that permute the original (but seemingly arbitrary) seeds given in Marsaglia with 64 bits of seed data. We simply generate pseudorandom, independent starting states for all data in all threads using a 32-bit MT19937 engine.