<cuda/std/random>

Provided functionalities

Random number engines:

• std::minstd_rand0

• std::minstd_rand

• C++26 std::philox4x32 - available from C++17 onwards

• C++26 std::philox4x64 - available from C++17 onwards

Note

cuda::pcg64 is provided in the non-standard <cuda/random> header. See cuda::pcg64.

Random number distributions:

• std::bernoulli_distribution

• std::binomial_distribution

• std::cauchy_distribution

• std::chi_squared_distribution

• std::exponential_distribution

• std::extreme_value_distribution

• std::fisher_f_distribution

• std::gamma_distribution

• std::geometric_distribution

• std::lognormal_distribution

• std::negative_binomial_distribution

• std::normal_distribution

• std::poisson_distribution

• std::student_t_distribution

• std::uniform_int_distribution

• std::uniform_real_distribution

• std::weibull_distribution

Utilities:

• std::seed_seq

• std::generate_canonical

Note

cuda::std::seed_seq should be used exclusively on host or exclusively on device. Do not share the same seed_seq instance between host and device code.

The following engines or distributions are not implemented as they are not convenient or practical to implement in CUDA device code, either due to dynamic memory allocations or large state sizes.

Not supported

• std::random_device

• std::mersenne_twister_engine (std::mt19937, std::mt19937_64)

• std::subtract_with_carry_engine (std::ranlux24_base, std::ranlux48_base, std::ranlux24, std::ranlux48)

• std::discard_block_engine

• std::independent_bits_engine

• std::shuffle_order_engine (std::knuth_b)

• std::discrete_distribution

• std::piecewise_constant_distribution

• std::piecewise_linear_distribution