StarPU Cholesky factorization

Functions
void	starpu_task_cpotrfsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, int prio)
	Submits starpu potrfsp cblk or blok task. More...
void	starpu_task_potrf_cgemmsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, const SolverBlok *blokB, SolverCblk *fcblk, int prio)
	Submits starpu cgemmsp cblk or blok task. More...
void	starpu_cpotrf_sp1dplus_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_cpotrf_sp1dplus_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_cpotrf_sp2d_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_cpotrf_sp2d_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_cpotrf (pastix_data_t *pastix_data, sopalin_data_t *sopalin_data)
	Perform a sparse Cholesky factorization using StarPU runtime. More...
void	starpu_task_zpotrfsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, int prio)
	Submits starpu potrfsp cblk or blok task. More...
void	starpu_task_potrf_zgemmsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, const SolverBlok *blokB, SolverCblk *fcblk, int prio)
	Submits starpu zgemmsp cblk or blok task. More...
void	starpu_zpotrf_sp1dplus_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_zpotrf_sp1dplus_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_zpotrf_sp2d_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_zpotrf_sp2d_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_zpotrf (pastix_data_t *pastix_data, sopalin_data_t *sopalin_data)
	Perform a sparse Cholesky factorization using StarPU runtime. More...
void	starpu_task_dpotrfsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, int prio)
	Submits starpu potrfsp cblk or blok task. More...
void	starpu_task_potrf_dgemmsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, const SolverBlok *blokB, SolverCblk *fcblk, int prio)
	Submits starpu dgemmsp cblk or blok task. More...
void	starpu_dpotrf_sp1dplus_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_dpotrf_sp1dplus_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_dpotrf_sp2d_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_dpotrf_sp2d_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_dpotrf (pastix_data_t *pastix_data, sopalin_data_t *sopalin_data)
	Perform a sparse Cholesky factorization using StarPU runtime. More...
void	starpu_task_spotrfsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, int prio)
	Submits starpu potrfsp cblk or blok task. More...
void	starpu_task_potrf_sgemmsp (sopalin_data_t *sopalin_data, SolverCblk *cblk, const SolverBlok *blokB, SolverCblk *fcblk, int prio)
	Submits starpu sgemmsp cblk or blok task. More...
void	starpu_spotrf_sp1dplus_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_spotrf_sp1dplus_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D kernels. More...
void	starpu_spotrf_sp2d_rl (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_spotrf_sp2d_ll (sopalin_data_t *sopalin_data, starpu_sparse_matrix_desc_t *desc)
	Perform a sparse Cholesky factorization with 1D and 2D kernels. More...
void	starpu_spotrf (pastix_data_t *pastix_data, sopalin_data_t *sopalin_data)
	Perform a sparse Cholesky factorization using StarPU runtime. More...

Detailed Description

Function Documentation

starpu_task_cpotrfsp()

void starpu_task_cpotrfsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		int	prio
	)

Submits starpu potrfsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	prio	The task priority.

Definition at line 45 of file starpu_cpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixConjTrans, PastixLCoef, PastixLower, PastixNonUnit, PastixRight, starpu_task_blok_cpotrf(), starpu_task_blok_ctrsmsp(), and starpu_task_cblk_cpotrfsp().

Referenced by starpu_cpotrf_sp2d_ll(), and starpu_cpotrf_sp2d_rl().

starpu_task_potrf_cgemmsp()

void starpu_task_potrf_cgemmsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		const SolverBlok *	blokB,
		SolverCblk *	fcblk,
		int	prio
	)

Submits starpu cgemmsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	blokB	The block.
[in]	fcblk	The facing column block.
[in]	prio	The task priority.

Definition at line 100 of file starpu_cpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, PastixConjTrans, PastixLCoef, starpu_task_blok_cgemmsp(), and starpu_task_cblk_cgemmsp().

Referenced by starpu_cpotrf_sp2d_ll(), and starpu_cpotrf_sp2d_rl().

starpu_cpotrf_sp1dplus_rl()

void starpu_cpotrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 154 of file starpu_cpotrf.c.

References solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixConjTrans, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cadd_1dp_fanin(), starpu_task_cadd_1dp_recv(), starpu_task_cblk_cgemmsp(), and starpu_task_cblk_cpotrfsp().

Referenced by starpu_cpotrf().

starpu_cpotrf_sp1dplus_ll()

void starpu_cpotrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 228 of file starpu_cpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::lcblknm, pastix_int_t, PastixConjTrans, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cadd_1dp_fanin(), starpu_task_cadd_1dp_recv(), starpu_task_cblk_cgemmsp(), and starpu_task_cblk_cpotrfsp().

Referenced by starpu_cpotrf().

starpu_cpotrf_sp2d_rl()

void starpu_cpotrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 309 of file starpu_cpotrf.c.

References solver_matrix_s::cblkmax1d, solver_matrix_s::cblkmin2d, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cadd_1dp_fanin(), starpu_task_cadd_1dp_recv(), starpu_task_cadd_2d_fanin(), starpu_task_cadd_2d_recv(), starpu_task_cpotrfsp(), and starpu_task_potrf_cgemmsp().

Referenced by starpu_cpotrf().

starpu_cpotrf_sp2d_ll()

void starpu_cpotrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 440 of file starpu_cpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::fcblknm, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cadd_fanin(), starpu_task_cadd_recv(), starpu_task_cpotrfsp(), and starpu_task_potrf_cgemmsp().

Referenced by starpu_cpotrf().

starpu_cpotrf()

void starpu_cpotrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse Cholesky factorization using StarPU runtime.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

The algorithm is automatically chosen between the 1D and 2D version based on the API parameter IPARM_TASKS2D_LEVEL. If IPARM_TASKS2D_LEVEL != 0 the 2D scheme is applied, the 1D otherwise.

Parameters

[in,out]	pastix_data	The pastix_data structure that describes the solver instance.
[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.

Definition at line 533 of file starpu_cpotrf.c.

References pastix_data_s::inter_node_comm, pastix_data_s::inter_node_procnbr, pastix_data_s::inter_node_procnum, pastix_data_s::iparm, IPARM_FACTO_LOOK_SIDE, IPARM_TASKS2D_LEVEL, IPARM_TRACE, pastix_data_s::pastix_comm, pastix_starpu_init(), PastixFactLeftLooking, PastixHermitian, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_cpotrf_sp1dplus_ll(), starpu_cpotrf_sp1dplus_rl(), starpu_cpotrf_sp2d_ll(), starpu_cpotrf_sp2d_rl(), starpu_sparse_matrix_getoncpu(), and starpu_sparse_matrix_init().

starpu_task_zpotrfsp()

void starpu_task_zpotrfsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		int	prio
	)

Submits starpu potrfsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	prio	The task priority.

Definition at line 45 of file starpu_zpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixConjTrans, PastixLCoef, PastixLower, PastixNonUnit, PastixRight, starpu_task_blok_zpotrf(), starpu_task_blok_ztrsmsp(), and starpu_task_cblk_zpotrfsp().

Referenced by starpu_zpotrf_sp2d_ll(), and starpu_zpotrf_sp2d_rl().

starpu_task_potrf_zgemmsp()

void starpu_task_potrf_zgemmsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		const SolverBlok *	blokB,
		SolverCblk *	fcblk,
		int	prio
	)

Submits starpu zgemmsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	blokB	The block.
[in]	fcblk	The facing column block.
[in]	prio	The task priority.

Definition at line 100 of file starpu_zpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, PastixConjTrans, PastixLCoef, starpu_task_blok_zgemmsp(), and starpu_task_cblk_zgemmsp().

Referenced by starpu_zpotrf_sp2d_ll(), and starpu_zpotrf_sp2d_rl().

starpu_zpotrf_sp1dplus_rl()

void starpu_zpotrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 154 of file starpu_zpotrf.c.

References solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixConjTrans, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cblk_zgemmsp(), starpu_task_cblk_zpotrfsp(), starpu_task_zadd_1dp_fanin(), and starpu_task_zadd_1dp_recv().

Referenced by starpu_zpotrf().

starpu_zpotrf_sp1dplus_ll()

void starpu_zpotrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 228 of file starpu_zpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::lcblknm, pastix_int_t, PastixConjTrans, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_cblk_zgemmsp(), starpu_task_cblk_zpotrfsp(), starpu_task_zadd_1dp_fanin(), and starpu_task_zadd_1dp_recv().

Referenced by starpu_zpotrf().

starpu_zpotrf_sp2d_rl()

void starpu_zpotrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 309 of file starpu_zpotrf.c.

References solver_matrix_s::cblkmax1d, solver_matrix_s::cblkmin2d, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_potrf_zgemmsp(), starpu_task_zadd_1dp_fanin(), starpu_task_zadd_1dp_recv(), starpu_task_zadd_2d_fanin(), starpu_task_zadd_2d_recv(), and starpu_task_zpotrfsp().

Referenced by starpu_zpotrf().

starpu_zpotrf_sp2d_ll()

void starpu_zpotrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 440 of file starpu_zpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::fcblknm, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_potrf_zgemmsp(), starpu_task_zadd_fanin(), starpu_task_zadd_recv(), and starpu_task_zpotrfsp().

Referenced by starpu_zpotrf().

starpu_zpotrf()

void starpu_zpotrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse Cholesky factorization using StarPU runtime.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Hermitian positive definite in the complex case) matrix A. The factorization has the form

\[ A = L\times L^H \]

where L is a sparse lower triangular matrix.

The algorithm is automatically chosen between the 1D and 2D version based on the API parameter IPARM_TASKS2D_LEVEL. If IPARM_TASKS2D_LEVEL != 0 the 2D scheme is applied, the 1D otherwise.

Parameters

[in,out]	pastix_data	The pastix_data structure that describes the solver instance.
[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.

Definition at line 533 of file starpu_zpotrf.c.

References pastix_data_s::inter_node_comm, pastix_data_s::inter_node_procnbr, pastix_data_s::inter_node_procnum, pastix_data_s::iparm, IPARM_FACTO_LOOK_SIDE, IPARM_TASKS2D_LEVEL, IPARM_TRACE, pastix_data_s::pastix_comm, pastix_starpu_init(), PastixFactLeftLooking, PastixHermitian, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_sparse_matrix_getoncpu(), starpu_sparse_matrix_init(), starpu_zpotrf_sp1dplus_ll(), starpu_zpotrf_sp1dplus_rl(), starpu_zpotrf_sp2d_ll(), and starpu_zpotrf_sp2d_rl().

starpu_task_dpotrfsp()

void starpu_task_dpotrfsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		int	prio
	)

Submits starpu potrfsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	prio	The task priority.

Definition at line 45 of file starpu_dpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixLower, PastixNonUnit, PastixRight, PastixTrans, starpu_task_blok_dpotrf(), starpu_task_blok_dtrsmsp(), and starpu_task_cblk_dpotrfsp().

Referenced by starpu_dpotrf_sp2d_ll(), and starpu_dpotrf_sp2d_rl().

starpu_task_potrf_dgemmsp()

void starpu_task_potrf_dgemmsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		const SolverBlok *	blokB,
		SolverCblk *	fcblk,
		int	prio
	)

Submits starpu dgemmsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	blokB	The block.
[in]	fcblk	The facing column block.
[in]	prio	The task priority.

Definition at line 100 of file starpu_dpotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, PastixLCoef, PastixTrans, starpu_task_blok_dgemmsp(), and starpu_task_cblk_dgemmsp().

Referenced by starpu_dpotrf_sp2d_ll(), and starpu_dpotrf_sp2d_rl().

starpu_dpotrf_sp1dplus_rl()

void starpu_dpotrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 154 of file starpu_dpotrf.c.

References solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, PastixTrans, starpu_sparse_cblk_wont_use(), starpu_task_cblk_dgemmsp(), starpu_task_cblk_dpotrfsp(), starpu_task_dadd_1dp_fanin(), and starpu_task_dadd_1dp_recv().

Referenced by starpu_dpotrf().

starpu_dpotrf_sp1dplus_ll()

void starpu_dpotrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 228 of file starpu_dpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, PastixTrans, starpu_sparse_cblk_wont_use(), starpu_task_cblk_dgemmsp(), starpu_task_cblk_dpotrfsp(), starpu_task_dadd_1dp_fanin(), and starpu_task_dadd_1dp_recv().

Referenced by starpu_dpotrf().

starpu_dpotrf_sp2d_rl()

void starpu_dpotrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 309 of file starpu_dpotrf.c.

References solver_matrix_s::cblkmax1d, solver_matrix_s::cblkmin2d, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_dadd_1dp_fanin(), starpu_task_dadd_1dp_recv(), starpu_task_dadd_2d_fanin(), starpu_task_dadd_2d_recv(), starpu_task_dpotrfsp(), and starpu_task_potrf_dgemmsp().

Referenced by starpu_dpotrf().

starpu_dpotrf_sp2d_ll()

void starpu_dpotrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 440 of file starpu_dpotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::fcblknm, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_dadd_fanin(), starpu_task_dadd_recv(), starpu_task_dpotrfsp(), and starpu_task_potrf_dgemmsp().

Referenced by starpu_dpotrf().

starpu_dpotrf()

void starpu_dpotrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse Cholesky factorization using StarPU runtime.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

The algorithm is automatically chosen between the 1D and 2D version based on the API parameter IPARM_TASKS2D_LEVEL. If IPARM_TASKS2D_LEVEL != 0 the 2D scheme is applied, the 1D otherwise.

Parameters

[in,out]	pastix_data	The pastix_data structure that describes the solver instance.
[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.

Definition at line 533 of file starpu_dpotrf.c.

References pastix_data_s::inter_node_comm, pastix_data_s::inter_node_procnbr, pastix_data_s::inter_node_procnum, pastix_data_s::iparm, IPARM_FACTO_LOOK_SIDE, IPARM_TASKS2D_LEVEL, IPARM_TRACE, pastix_data_s::pastix_comm, pastix_starpu_init(), PastixFactLeftLooking, PastixSymmetric, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_dpotrf_sp1dplus_ll(), starpu_dpotrf_sp1dplus_rl(), starpu_dpotrf_sp2d_ll(), starpu_dpotrf_sp2d_rl(), starpu_sparse_matrix_getoncpu(), and starpu_sparse_matrix_init().

starpu_task_spotrfsp()

void starpu_task_spotrfsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		int	prio
	)

Submits starpu potrfsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	prio	The task priority.

Definition at line 45 of file starpu_spotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixLower, PastixNonUnit, PastixRight, PastixTrans, starpu_task_blok_spotrf(), starpu_task_blok_strsmsp(), and starpu_task_cblk_spotrfsp().

Referenced by starpu_spotrf_sp2d_ll(), and starpu_spotrf_sp2d_rl().

starpu_task_potrf_sgemmsp()

void starpu_task_potrf_sgemmsp	(	sopalin_data_t *	sopalin_data,
		SolverCblk *	cblk,
		const SolverBlok *	blokB,
		SolverCblk *	fcblk,
		int	prio
	)

Submits starpu sgemmsp cblk or blok task.

Parameters

[in]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in]	cblk	The column block.
[in]	blokB	The block.
[in]	fcblk	The facing column block.
[in]	prio	The task priority.

Definition at line 100 of file starpu_spotrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, PastixLCoef, PastixTrans, starpu_task_blok_sgemmsp(), and starpu_task_cblk_sgemmsp().

Referenced by starpu_spotrf_sp2d_ll(), and starpu_spotrf_sp2d_rl().

starpu_spotrf_sp1dplus_rl()

void starpu_spotrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 154 of file starpu_spotrf.c.

References solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, PastixTrans, starpu_sparse_cblk_wont_use(), starpu_task_cblk_sgemmsp(), starpu_task_cblk_spotrfsp(), starpu_task_sadd_1dp_fanin(), and starpu_task_sadd_1dp_recv().

Referenced by starpu_spotrf().

starpu_spotrf_sp1dplus_ll()

void starpu_spotrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 228 of file starpu_spotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, PastixTrans, starpu_sparse_cblk_wont_use(), starpu_task_cblk_sgemmsp(), starpu_task_cblk_spotrfsp(), starpu_task_sadd_1dp_fanin(), and starpu_task_sadd_1dp_recv().

Referenced by starpu_spotrf().

starpu_spotrf_sp2d_rl()

void starpu_spotrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 309 of file starpu_spotrf.c.

References solver_matrix_s::cblkmax1d, solver_matrix_s::cblkmin2d, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_cblk_s::fblokptr, solver_blok_s::fcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_potrf_sgemmsp(), starpu_task_sadd_1dp_fanin(), starpu_task_sadd_1dp_recv(), starpu_task_sadd_2d_fanin(), starpu_task_sadd_2d_recv(), and starpu_task_spotrfsp().

Referenced by starpu_spotrf().

starpu_spotrf_sp2d_ll()

void starpu_spotrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse Cholesky factorization with 1D and 2D kernels.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

Parameters

[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.
[in,out]	desc	StarPU descriptor of the sparse matrix.

Definition at line 440 of file starpu_spotrf.c.

References solver_matrix_s::bloktab, solver_cblk_s::brownum, solver_matrix_s::browtab, solver_matrix_s::cblknbr, solver_matrix_s::cblktab, solver_cblk_s::cblktype, solver_blok_s::fcblknm, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, starpu_sparse_cblk_wont_use(), starpu_task_potrf_sgemmsp(), starpu_task_sadd_fanin(), starpu_task_sadd_recv(), and starpu_task_spotrfsp().

Referenced by starpu_spotrf().

starpu_spotrf()

void starpu_spotrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse Cholesky factorization using StarPU runtime.

The function performs the Cholesky factorization of a sparse symmetric positive definite (or Symmetric positive definite in the real case) matrix A. The factorization has the form

\[ A = L\times L^T \]

where L is a sparse lower triangular matrix.

The algorithm is automatically chosen between the 1D and 2D version based on the API parameter IPARM_TASKS2D_LEVEL. If IPARM_TASKS2D_LEVEL != 0 the 2D scheme is applied, the 1D otherwise.

Parameters

[in,out]	pastix_data	The pastix_data structure that describes the solver instance.
[in,out]	sopalin_data	Solver matrix information structure that will guide the algorithm.

Definition at line 533 of file starpu_spotrf.c.

References pastix_data_s::inter_node_comm, pastix_data_s::inter_node_procnbr, pastix_data_s::inter_node_procnum, pastix_data_s::iparm, IPARM_FACTO_LOOK_SIDE, IPARM_TASKS2D_LEVEL, IPARM_TRACE, pastix_data_s::pastix_comm, pastix_starpu_init(), PastixFactLeftLooking, PastixSymmetric, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_sparse_matrix_getoncpu(), starpu_sparse_matrix_init(), starpu_spotrf_sp1dplus_ll(), starpu_spotrf_sp1dplus_rl(), starpu_spotrf_sp2d_ll(), and starpu_spotrf_sp2d_rl().