gko::matrix::Csr< ValueType, IndexType >::automatical Class Reference

Inheritance diagram for gko::matrix::Csr< ValueType, IndexType >::automatical:

Collaboration diagram for gko::matrix::Csr< ValueType, IndexType >::automatical:

Public Member Functions
	automatical ()
	Creates an automatical strategy.
	automatical (std::shared_ptr< const CudaExecutor > exec)
	Creates an automatical strategy with CUDA executor.
	automatical (std::shared_ptr< const HipExecutor > exec)
	Creates an automatical strategy with HIP executor.
	automatical (std::shared_ptr< const DpcppExecutor > exec)
	Creates an automatical strategy with Dpcpp executor.
	automatical (int64_t nwarps, int warp_size=32, bool cuda_strategy=true, std::string strategy_name="none")
	Creates an automatical strategy with specified parameters.
void	process (const array< index_type > &mtx_row_ptrs, array< index_type > *mtx_srow) override
	Computes srow according to row pointers.
int64_t	clac_size (const int64_t nnz) override
	Computes the srow size according to the number of nonzeros.
index_type	get_max_length_per_row () const noexcept
std::shared_ptr< strategy_type >	copy () override
	Copy a strategy.
Public Member Functions inherited from gko::matrix::Csr< ValueType, IndexType >::strategy_type
	strategy_type (std::string name)
	Creates a strategy_type.
std::string	get_name ()
	Returns the name of strategy.

Public Attributes
const index_type	nvidia_row_len_limit = 1024
const index_type	nvidia_nnz_limit {static_cast<index_type>(1e6)}
const index_type	amd_row_len_limit = 768
const index_type	amd_nnz_limit {static_cast<index_type>(1e8)}
const index_type	intel_row_len_limit = 25600
const index_type	intel_nnz_limit {static_cast<index_type>(3e8)}

Constructor & Destructor Documentation

automatical() [1/5]

template<typename ValueType = default_precision, typename IndexType = int32>

gko::matrix::Csr< ValueType, IndexType >::automatical::automatical ( )

inline

Creates an automatical strategy.

Warning

this is deprecated! Please rely on the new automatic strategy instantiation or use one of the other constructors.

References automatical(), and gko::matrix::Csr< ValueType, IndexType >::create().

Referenced by automatical(), automatical(), automatical(), and automatical().

automatical() [2/5]

template<typename ValueType = default_precision, typename IndexType = int32>

gko::matrix::Csr< ValueType, IndexType >::automatical::automatical ( std::shared_ptr< const CudaExecutor > exec )

inline

Creates an automatical strategy with CUDA executor.

Parameters

exec	the CUDA executor

References automatical().

automatical() [3/5]

template<typename ValueType = default_precision, typename IndexType = int32>

gko::matrix::Csr< ValueType, IndexType >::automatical::automatical ( std::shared_ptr< const HipExecutor > exec )

inline

Creates an automatical strategy with HIP executor.

Parameters

exec	the HIP executor

References automatical().

automatical() [4/5]

template<typename ValueType = default_precision, typename IndexType = int32>

gko::matrix::Csr< ValueType, IndexType >::automatical::automatical ( std::shared_ptr< const DpcppExecutor > exec )

inline

Creates an automatical strategy with Dpcpp executor.

Parameters

exec	the Dpcpp executor

Note

TODO: porting - we hardcode the subgroup size is 32

References automatical().

automatical() [5/5]

template<typename ValueType = default_precision, typename IndexType = int32>

gko::matrix::Csr< ValueType, IndexType >::automatical::automatical ( int64_t nwarps, int warp_size = 32, bool cuda_strategy = true, std::string strategy_name = "none" )

inline

Creates an automatical strategy with specified parameters.

Parameters

nwarps	the number of warps in the executor
warp_size	the warp size of the executor
cuda_strategy	whether the cuda_strategy needs to be used.

Note

The warp_size must be the size of full warp. When using this constructor, set_strategy needs to be called with correct parameters which is replaced during the conversion.

References gko::matrix::Csr< ValueType, IndexType >::strategy_type::strategy_type().

Member Function Documentation

clac_size()

template<typename ValueType = default_precision, typename IndexType = int32>

int64_t gko::matrix::Csr< ValueType, IndexType >::automatical::clac_size ( const int64_t nnz )

inlineoverridevirtual

Computes the srow size according to the number of nonzeros.

Parameters

nnz	the number of nonzeros

Returns

the size of srow

Implements gko::matrix::Csr< ValueType, IndexType >::strategy_type.

copy()

template<typename ValueType = default_precision, typename IndexType = int32>

std::shared_ptr< strategy_type > gko::matrix::Csr< ValueType, IndexType >::automatical::copy ( )

inlineoverridevirtual

Copy a strategy.

This is a workaround until strategies are revamped, since strategies like automatical do not work when actually shared.

Implements gko::matrix::Csr< ValueType, IndexType >::strategy_type.

process()

template<typename ValueType = default_precision, typename IndexType = int32>

void gko::matrix::Csr< ValueType, IndexType >::automatical::process ( const array< index_type > & mtx_row_ptrs, array< index_type > * mtx_srow )

inlineoverridevirtual

Computes srow according to row pointers.

Parameters

mtx_row_ptrs	the row pointers of the matrix
mtx_srow	the srow of the matrix

Implements gko::matrix::Csr< ValueType, IndexType >::strategy_type.

References gko::array< ValueType >::get_const_data(), gko::array< ValueType >::get_executor(), gko::matrix::Csr< ValueType, IndexType >::strategy_type::get_name(), gko::array< ValueType >::get_size(), gko::matrix::Csr< ValueType, IndexType >::classical::process(), and gko::matrix::Csr< ValueType, IndexType >::load_balance::process().

---

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

• ginkgo/core/matrix/csr.hpp