ordered_logistic()

__kernel void stan::math::opencl_kernels::ordered_logistic	(	__global double *	logp_global,
		__global double *	lambda_derivative,
		__global double *	cuts_derivative,
		const __global int *	y_global,
		const __global double *	lambda_global,
		const __global double *	cuts,
		const int	N_instances,
		const int	N_classes,
		const int	is_y_vector,
		const int	is_cuts_matrix,
		const int	need_lambda_derivative,
		const int	need_cuts_derivative
	)

GPU implementation of ordinal regression.

Must be run with at least N_instances threads and local size equal to LOCAL_SIZE_.

Parameters

[out]	logp_global	partially summed log probability (1 value per work group)
[out]	lambda_derivative	derivative wrt lambda
[out]	cuts_derivative	partially summed derivative wrt cuts (1 column per work group)
[in]	y_global	a scalar or vector of classes.
[in]	lambda_global	vector of continuous lambda variables
[in]	cuts	cutpoints vector
	N_instances	number of cases
	N_classes	number of classes
	is_y_vector	0 or 1 - whether y is a vector (alternatively it is a scalar we need to broadcast)
	is_cuts_matrix	0 or 1 - whether cuts is a matrix (alternatively it is a vector we need to broadcast)
	need_lambda_derivative	0 or 1 - whether lambda_derivative needs to be computed
	need_cuts_derivative	0 or 1 - whether cuts_derivative needs to be computed

Definition at line 41 of file ordered_logistic_lpmf.hpp.