pcg64_dxsm

pcg64_dxsm#

A permuted congruential pseudorandom number generator PCG64 DXSM with a period of \(2^{128}\) [pcg2014].

Description

The PCG is a family of random number generators based on linear congruential generator (LCG). The algorithm applies permutation function to improve statistical properties of LCG. PCG64 DXSM uses double xorshift multiply output function and has a period of \(2^{128}\).

Generation algorithm

\(x_n = x_{n-1} \cdot a + b\)

\(hi_n = x_n \gg 64\)

\(lo_n = x_n \land \left(2^{64} - 1\right) \lor 1\)

\(hi_n = hi_n \oplus (hi_n \gg 32)\)

\(hi_n = hi_n \cdot a\)

\(hi_n = hi_n \oplus (hi_n \gg 48)\)

\(u_n = hi_n \cdot lo_n\)

\(a = \text{0xDA942042E4DD58B5}, b = \text{0x5851F42D4C957F2D14057B7EF767814F}\)

class pcg64_dxsm#

Syntax

namespace oneapi::math::rng::device {
    template<std::int32_t VecSize = 1>
    class pcg64_dxsm {
    public:
        static constexpr std::uint64_t default_seed = 1;
        static constexpr std::int32_t vec_size = VecSize;

        pcg64_dxsm();
        pcg64_dxsm(std::uint64_t seed, std::uint64_t offset = 0);
        pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::uint64_t offset = 0);
        pcg64_dxsm(std::uint64_t seed, std::initializer_list<std::uint64_t> offset);
        pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::initializer_list<std::uint64_t> offset);
    };
}

Class Template Parameters

VecSize: Describes the size of vector which will be produced by generate function by this engine. VecSize values may be 1, 2, 3, 4, 8, 16 as sycl::vec class size. By default VecSize = 1, for this case, a single random number is returned by the generate function.

Class Members

Routine	Description
pcg64_dxsm()	Default constructor
pcg64_dxsm(std::uint64_t seed, std::uint64_t offset = 0)	Constructor for common seed initialization of the engine and common number of skipped elements
pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::uint64_t offset = 0)	Constructor for extended seed initialization of the engine and common number of skipped elements
pcg64_dxsm(std::uint64_t seed, std::initializer_list<std::uint64_t> offset)	Constructor for common seed initialization of the engine and extended number of skipped elements
pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::initializer_list<std::uint64_t> offset)	Constructor for extended seed initialization of the engine and extended number of skipped elements

Constructors

pcg64_dxsm::pcg64_dxsm()

pcg64_dxsm::pcg64_dxsm(std::uint64_t seed, std::uint64_t offset = 0)

Input Parameters

seed: The initial conditions of the generator state.
offset: Number of skipped elements.

pcg64_dxsm::pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::uint64_t offset = 0)

Input Parameters

seed: The initial conditions of the generator state.
offset: Number of skipped elements.

pcg64_dxsm::pcg64_dxsm(std::uint64_t seed, std::initializer_list<std::uint64_t> offset)

Input Parameters

seed: The initial conditions of the generator state.
offset: Number of skipped elements. Offset is calculated as: num_to_skip [0]+ num_to_skip [1]*2⁶⁴.

pcg64_dxsm::pcg64_dxsm(std::initializer_list<std::uint64_t> seed, std::initializer_list<std::uint64_t> offset)

Input Parameters

seed: The initial conditions of the generator state.
offset: Number of skipped elements. Offset is calculated as: num_to_skip [0]+ num_to_skip [1]*2⁶⁴.

Parent topic: Device Engines (Basic Random Number Generators)

pcg64_dxsm

Contents

pcg64_dxsm#

class pcg64_dxsm#