CUDA Wrappers

Classes
class	CUDAWrappers::SolverDirect< Number >
class	CUDAWrappers::FEEvaluation< dim, fe_degree, n_q_points_1d, n_components_, Number >
class	CUDAWrappers::MatrixFree< dim, Number >
struct	CUDAWrappers::internal::EvaluatorTensorProduct< variant, dim, fe_degree, n_q_points_1d, Number >
struct	CUDAWrappers::internal::EvaluatorTensorProduct< evaluate_general, dim, fe_degree, n_q_points_1d, Number >

Enumerations

Functions
float	LinearAlgebra::CUDAWrappers::atomicAdd_wrapper (float *address, float val)
double	LinearAlgebra::CUDAWrappers::atomicAdd_wrapper (double *address, double val)
float	LinearAlgebra::CUDAWrappers::atomicMax_wrapper (float *address, float val)
double	LinearAlgebra::CUDAWrappers::atomicMax_wrapper (double *address, double val)

Detailed Description

The classes in this module are concerned with the description of features to be run on GPUs using CUDA.

Enumeration Type Documentation

enum CUDAWrappers::internal::EvaluatorVariant

In this namespace, the evaluator routines that evaluate the tensor products are implemented.

Definition at line 37 of file cuda_tensor_product_kernels.h.

Function Documentation

float LinearAlgebra::CUDAWrappers::atomicAdd_wrapper ( float *  address, float  val  )

inline

Provide atomicAdd for floats.

Definition at line 35 of file cuda_atomic.h.

double LinearAlgebra::CUDAWrappers::atomicAdd_wrapper ( double *  address, double  val  )

inline

Provide atomicAdd for doubles.

Definition at line 48 of file cuda_atomic.h.

float LinearAlgebra::CUDAWrappers::atomicMax_wrapper ( float *  address, float  val  )

inline

Provide atomicMax for floats.

Definition at line 81 of file cuda_atomic.h.

double LinearAlgebra::CUDAWrappers::atomicMax_wrapper ( double *  address, double  val  )

inline

Provide atomicMax for doubles.

Definition at line 105 of file cuda_atomic.h.