Functions
void	batchQRFactorize (ParallelManager pm, double P, int lda, int nda, double RHS, int ldb, int ndb, int M, int N, int NRHS, const int num_matrices, const size_t max_neighbors=0, const int initial_index_of_batch=0, int *neighbor_list_sizes=NULL)
	Calls LAPACK or CUBLAS to solve a batch of QR problems. More...

void	batchSVDFactorize (ParallelManager pm, bool swap_layout_P, double P, int lda, int nda, bool swap_layout_RHS, double RHS, int ldb, int ndb, int M, int N, int NRHS, const int num_matrices, const size_t max_neighbors=0, const int initial_index_of_batch=0, int *neighbor_list_sizes=NULL)
	Calls LAPACK or CUBLAS to solve a batch of SVD problems. More...

void	batchLUFactorize (ParallelManager pm, double P, int lda, int nda, double RHS, int ldb, int ndb, int M, int N, int NRHS, const int num_matrices, const size_t max_neighbors=0, const int initial_index_of_batch=0, int *neighbor_list_sizes=NULL)
	Calls LAPACK or CUBLAS to solve a batch of LU problems. More...

KOKKOS_INLINE_FUNCTION void	createM (const member_type &teamMember, scratch_matrix_right_type M_data, scratch_matrix_right_type weighted_P, const int columns, const int rows)
	Creates a matrix M=A^T*A from a matrix A. More...

KOKKOS_INLINE_FUNCTION void	largestTwoEigenvectorsThreeByThreeSymmetric (const member_type &teamMember, scratch_matrix_right_type V, scratch_matrix_right_type PtP, const int dimensions, pool_type &random_number_pool)
	Calculates two eigenvectors corresponding to two dominant eigenvalues. More...

Function Documentation

void Compadre::GMLS_LinearAlgebra::batchLUFactorize	(	ParallelManager	pm,
		double *	P,
		int	lda,
		int	nda,
		double *	RHS,
		int	ldb,
		int	ndb,
		int	M,
		int	N,
		int	NRHS,
		const int	num_matrices,
		const size_t	max_neighbors = `0`,
		const int	initial_index_of_batch = `0`,
		int *	neighbor_list_sizes = `NULL`
	)

Calls LAPACK or CUBLAS to solve a batch of LU problems.

P contains num_matrices * lda * ndb data which is num_matrices different matrices, and RHS contains num_matrices * ldn * ndb data which is num_matrices different matrix right hand sides.

Parameters

pm	[in] - manager class for team and thread parallelism
P	[in/out] - evaluation of sampling functional on polynomial basis (in), meaningless workspace output (out)
lda	[in] - row dimension of each matrix in P
nda	[in] - columns dimension of each matrix in P
RHS	[in/out] - basis to invert P against (in), polynomial coefficients (out)
ldb	[in] - row dimension of each matrix in RHS
ndb	[in] - column dimension of each matrix in RHS
M	[in] - number of rows containing data (maximum rows over everything in batch)
N	[in] - number of columns containing data
num_matrices	[in] - number of target (GMLS problems to be solved)
max_neighbors	[in] - integer for maximum neighbor over all targets
neighbor_list_sizes	[in] - pointer to all neighbor list sizes for each target

Definition at line 581 of file Compadre_LinearAlgebra.cpp.

void Compadre::GMLS_LinearAlgebra::batchQRFactorize	(	ParallelManager	pm,
		double *	P,
		int	lda,
		int	nda,
		double *	RHS,
		int	ldb,
		int	ndb,
		int	M,
		int	N,
		int	NRHS,
		const int	num_matrices,
		const size_t	max_neighbors = `0`,
		const int	initial_index_of_batch = `0`,
		int *	neighbor_list_sizes = `NULL`
	)

Calls LAPACK or CUBLAS to solve a batch of QR problems.

P contains num_matrices * lda * ndb data which is num_matrices different matrices, and RHS contains num_matrices * ldn * ndb data which is num_matrices different matrix right hand sides.

Parameters

pm	[in] - manager class for team and thread parallelism
P	[in/out] - evaluation of sampling functional on polynomial basis (in), meaningless workspace output (out)
lda	[in] - row dimension of each matrix in P
nda	[in] - columns dimension of each matrix in P
RHS	[in/out] - basis to invert P against (in), polynomial coefficients (out)
ldb	[in] - row dimension of each matrix in RHS
ndb	[in] - column dimension of each matrix in RHS
M	[in] - number of rows containing data (maximum rows over everything in batch)
N	[in] - number of columns containing data
num_matrices	[in] - number of target (GMLS problems to be solved)
max_neighbors	[in] - integer for maximum neighbor over all targets
neighbor_list_sizes	[in] - pointer to all neighbor list sizes for each target

Definition at line 14 of file Compadre_LinearAlgebra.cpp.

void Compadre::GMLS_LinearAlgebra::batchSVDFactorize	(	ParallelManager	pm,
		bool	swap_layout_P,
		double *	P,
		int	lda,
		int	nda,
		bool	swap_layout_RHS,
		double *	RHS,
		int	ldb,
		int	ndb,
		int	M,
		int	N,
		int	NRHS,
		const int	num_matrices,
		const size_t	max_neighbors = `0`,
		const int	initial_index_of_batch = `0`,
		int *	neighbor_list_sizes = `NULL`
	)

Calls LAPACK or CUBLAS to solve a batch of SVD problems.

P contains num_matrices * lda * ndb data which is num_matrices different matrices, and RHS contains num_matrices * ldn * ndb data which is num_matrices different matrix right hand sides.

Parameters

pm	[in] - manager class for team and thread parallelism
swap_layout_P	[in] - boolean to check if layout of P will need to be swapped or not
P	[in/out] - evaluation of sampling functional on polynomial basis (in), meaningless workspace output (out)
lda	[in] - row dimension of each matrix in P
nda	[in] - columns dimension of each matrix in P
swap_layout_RHS	[in] - boolean to check if layout of RHS will need to be swapped or not
RHS	[in/out] - basis to invert P against (in), polynomial coefficients (out)
ldb	[in] - row dimension of each matrix in RHS
ndb	[in] - column dimension of each matrix in RHS
M	[in] - number of rows containing data (maximum rows over everything in batch)
N	[in] - number of columns containing data
num_matrices	[in] - number of target (GMLS problems to be solved)
max_neighbors	[in] - integer for maximum neighbor over all targets
neighbor_list_sizes	[in] - pointer to all neighbor list sizes for each target

Definition at line 164 of file Compadre_LinearAlgebra.cpp.

KOKKOS_INLINE_FUNCTION void Compadre::GMLS_LinearAlgebra::createM	(	const member_type &	teamMember,
		scratch_matrix_right_type	M_data,
		scratch_matrix_right_type	weighted_P,
		const int	columns,
		const int	rows
	)

Creates a matrix M=A^T*A from a matrix A.

Parameters

teamMember	[in] - Kokkos::TeamPolicy member type (created by parallel_for)
M_data	[out] - result of weighted_P^T * weighted_P
weighted_P	[in] - matrix to be multiplied by its transpose
columns	[in] - number of columns in weighted_P
rows	[in] - number of rows in weighted_P

Definition at line 10 of file Compadre_LinearAlgebra_Definitions.hpp.

KOKKOS_INLINE_FUNCTION void Compadre::GMLS_LinearAlgebra::largestTwoEigenvectorsThreeByThreeSymmetric	(	const member_type &	teamMember,
		scratch_matrix_right_type	V,
		scratch_matrix_right_type	PtP,
		const int	dimensions,
		pool_type &	random_number_pool
	)

Calculates two eigenvectors corresponding to two dominant eigenvalues.

Parameters

teamMember	[in] - Kokkos::TeamPolicy member type (created by parallel_for)
V	[out] - dimension * dimension Kokkos View
PtP	[in] - dimension * dimension Kokkos View
dimensions	[in] - dimension of PtP

Definition at line 55 of file Compadre_LinearAlgebra_Definitions.hpp.

Functions

Function Documentation