template class xf::fintech::MultiVariateNormalRng¶

#include "rng.hpp"

Overview¶

Multi-variate normal distribution RNG.

Parameters:

_DT	data type, either float or double.
_VariateNum	number of variates.
_BuffDepth	depth of buffer.

template <
    typename _DT,
    int _VariatePairNum,
    int _BuffDepth
    >
class MultiVariateNormalRng

MultiVariateNormalRng ()

Constructor, use pragma to determine data member’s hardware implementation.

void init (
    ap_uint <32> seed,
    _DT input_ltm [_VariatePairNum *2+1][_VariatePairNum]
    )

Initialize underlying RNG, setup lower triangle matrix and pre-calculate.

Parameters:

seed	seed to initialize underlying RNG
input_ltm	input lower triangle matrix

void next (
    _DT& res_0,
    _DT& res_1
    )

Each call returns two random number, in the order of 1st and 2nd, 3rd and 4th …. 2*n-1 th and 2*n th.

Parameters:

res_0	(2*n - 1)-th random number generated.
res_1	2*n-th random number generated.