Reference documentation for deal.II version Git f6af164dc0 2020-01-19 13:12:56 +0100
\(\newcommand{\vcentcolon}{\mathrel{\mathop{:}}}\) \(\newcommand{\dealcoloneq}{\vcentcolon\mathrel{\mkern-1.2mu}=}\) \(\newcommand{\jump}[1]{\left[\!\left[ #1 \right]\!\right]}\) \(\newcommand{\average}[1]{\left\{\!\left\{ #1 \right\}\!\right\}}\)
Public Member Functions | Public Attributes | List of all members
CUDAWrappers::PreconditionIC< Number >::AdditionalData Struct Reference

#include <deal.II/lac/cuda_precondition.h>

Public Member Functions

 AdditionalData (bool use_level_analysis=true)
 

Public Attributes

bool use_level_analysis
 

Detailed Description

template<typename Number>
struct CUDAWrappers::PreconditionIC< Number >::AdditionalData

Standardized data struct to pipe additional flags to the preconditioner.

Definition at line 76 of file cuda_precondition.h.

Constructor & Destructor Documentation

◆ AdditionalData()

template<typename Number >
CUDAWrappers::PreconditionIC< Number >::AdditionalData::AdditionalData ( bool  use_level_analysis = true)

Constructor. cuSPARSE allows to compute and use level information. According to the documentation this might improve performance. It is suggested to try both options.

Member Data Documentation

◆ use_level_analysis

template<typename Number >
bool CUDAWrappers::PreconditionIC< Number >::AdditionalData::use_level_analysis

Flag that determines if level information is used when creating and applying the preconditioner. See the documentation for cusparseSolvePolicy_t at https://docs.nvidia.com/cuda/cusparse/index.html#cusparsesolvepolicy_t for more information.

Definition at line 92 of file cuda_precondition.h.


The documentation for this struct was generated from the following file: