PaStiX Handbook: StarPU LU factorization

Functions
void	starpu_task_getrf_zgemmsp (sopalin_data_t sopalin_data, SolverCblk cblk, const SolverBlok blokB, SolverCblk fcblk, int prio)
	TODO. More...

void	starpu_task_zgetrfsp (sopalin_data_t sopalin_data, SolverCblk cblk, int prio)
	TODO. More...

void	starpu_zgetrf_sp1dplus_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_zgetrf_sp1dplus_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_zgetrf_sp2d_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_zgetrf_sp2d_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_zgetrf (pastix_data_t pastix_data, sopalin_data_t sopalin_data)
	Perform a sparse LU factorization using StarPU runtime. More...

void	starpu_task_getrf_cgemmsp (sopalin_data_t sopalin_data, SolverCblk cblk, const SolverBlok blokB, SolverCblk fcblk, int prio)
	TODO. More...

void	starpu_task_cgetrfsp (sopalin_data_t sopalin_data, SolverCblk cblk, int prio)
	TODO. More...

void	starpu_cgetrf_sp1dplus_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_cgetrf_sp1dplus_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_cgetrf_sp2d_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_cgetrf_sp2d_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_cgetrf (pastix_data_t pastix_data, sopalin_data_t sopalin_data)
	Perform a sparse LU factorization using StarPU runtime. More...

void	starpu_task_getrf_dgemmsp (sopalin_data_t sopalin_data, SolverCblk cblk, const SolverBlok blokB, SolverCblk fcblk, int prio)
	TODO. More...

void	starpu_task_dgetrfsp (sopalin_data_t sopalin_data, SolverCblk cblk, int prio)
	TODO. More...

void	starpu_dgetrf_sp1dplus_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_dgetrf_sp1dplus_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_dgetrf_sp2d_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_dgetrf_sp2d_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_dgetrf (pastix_data_t pastix_data, sopalin_data_t sopalin_data)
	Perform a sparse LU factorization using StarPU runtime. More...

void	starpu_task_getrf_sgemmsp (sopalin_data_t sopalin_data, SolverCblk cblk, const SolverBlok blokB, SolverCblk fcblk, int prio)
	TODO. More...

void	starpu_task_sgetrfsp (sopalin_data_t sopalin_data, SolverCblk cblk, int prio)
	TODO. More...

void	starpu_sgetrf_sp1dplus_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_sgetrf_sp1dplus_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D kernels. More...

void	starpu_sgetrf_sp2d_rl (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_sgetrf_sp2d_ll (sopalin_data_t sopalin_data, starpu_sparse_matrix_desc_t desc)
	Perform a sparse LU factorization with 1D and 2D kernels. More...

void	starpu_sgetrf (pastix_data_t pastix_data, sopalin_data_t sopalin_data)
	Perform a sparse LU factorization using StarPU runtime. More...

Detailed Description

Function Documentation

◆ starpu_task_getrf_zgemmsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	fcblk	TODO
[in]	blokA	TODO
[in]	prio	TODO

Definition at line 55 of file starpu_zgetrf.c.

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

Referenced by starpu_zgetrf_sp2d_ll(), and starpu_zgetrf_sp2d_rl().

◆ starpu_task_zgetrfsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	prio	TODO

Definition at line 125 of file starpu_zgetrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixNonUnit, PastixNoTrans, PastixRight, PastixUCoef, PastixUnit, PastixUpper, starpu_task_blok_zgetrf(), starpu_task_blok_ztrsmsp(), and starpu_task_cblk_zgetrfsp().

Referenced by starpu_zgetrf_sp2d_ll(), and starpu_zgetrf_sp2d_rl().

◆ starpu_zgetrf_sp1dplus_rl()

void starpu_zgetrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 183 of file starpu_zgetrf.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, PastixLUCoef, PastixTrans, PastixUCoef, starpu_sparse_cblk_wont_use(), starpu_task_cblk_zgemmsp(), and starpu_task_cblk_zgetrfsp().

Referenced by starpu_zgetrf().

◆ starpu_zgetrf_sp1dplus_ll()

void starpu_zgetrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 251 of file starpu_zgetrf.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_cblk_s::fblokptr, solver_blok_s::fcblknm, solver_blok_s::lcblknm, pastix_int_t, PastixLCoef, PastixTrans, PastixUCoef, starpu_sparse_cblk_wont_use(), starpu_task_cblk_zgemmsp(), and starpu_task_cblk_zgetrfsp().

Referenced by starpu_zgetrf().

◆ starpu_zgetrf_sp2d_rl()

void starpu_zgetrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 329 of file starpu_zgetrf.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, PastixLUCoef, starpu_sparse_cblk_wont_use(), starpu_task_getrf_zgemmsp(), and starpu_task_zgetrfsp().

Referenced by starpu_zgetrf().

◆ starpu_zgetrf_sp2d_ll()

void starpu_zgetrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 428 of file starpu_zgetrf.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, PastixLUCoef, starpu_sparse_cblk_wont_use(), starpu_task_getrf_zgemmsp(), and starpu_task_zgetrfsp().

Referenced by starpu_zgetrf().

◆ starpu_zgetrf()

void starpu_zgetrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse LU factorization using StarPU runtime.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 505 of file starpu_zgetrf.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, PastixGeneral, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_sparse_matrix_getoncpu(), starpu_sparse_matrix_init(), starpu_zgetrf_sp1dplus_ll(), starpu_zgetrf_sp1dplus_rl(), starpu_zgetrf_sp2d_ll(), and starpu_zgetrf_sp2d_rl().

◆ starpu_task_getrf_cgemmsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	fcblk	TODO
[in]	blokA	TODO
[in]	prio	TODO

Definition at line 55 of file starpu_cgetrf.c.

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

Referenced by starpu_cgetrf_sp2d_ll(), and starpu_cgetrf_sp2d_rl().

◆ starpu_task_cgetrfsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	prio	TODO

Definition at line 125 of file starpu_cgetrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixNonUnit, PastixNoTrans, PastixRight, PastixUCoef, PastixUnit, PastixUpper, starpu_task_blok_cgetrf(), starpu_task_blok_ctrsmsp(), and starpu_task_cblk_cgetrfsp().

Referenced by starpu_cgetrf_sp2d_ll(), and starpu_cgetrf_sp2d_rl().

◆ starpu_cgetrf_sp1dplus_rl()

void starpu_cgetrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 183 of file starpu_cgetrf.c.

Referenced by starpu_cgetrf().

◆ starpu_cgetrf_sp1dplus_ll()

void starpu_cgetrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 251 of file starpu_cgetrf.c.

Referenced by starpu_cgetrf().

◆ starpu_cgetrf_sp2d_rl()

void starpu_cgetrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 329 of file starpu_cgetrf.c.

Referenced by starpu_cgetrf().

◆ starpu_cgetrf_sp2d_ll()

void starpu_cgetrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 428 of file starpu_cgetrf.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, PastixLUCoef, starpu_sparse_cblk_wont_use(), starpu_task_cgetrfsp(), and starpu_task_getrf_cgemmsp().

Referenced by starpu_cgetrf().

◆ starpu_cgetrf()

void starpu_cgetrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse LU factorization using StarPU runtime.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 505 of file starpu_cgetrf.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, PastixGeneral, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_cgetrf_sp1dplus_ll(), starpu_cgetrf_sp1dplus_rl(), starpu_cgetrf_sp2d_ll(), starpu_cgetrf_sp2d_rl(), starpu_sparse_matrix_getoncpu(), and starpu_sparse_matrix_init().

◆ starpu_task_getrf_dgemmsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	fcblk	TODO
[in]	blokA	TODO
[in]	prio	TODO

Definition at line 55 of file starpu_dgetrf.c.

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

Referenced by starpu_dgetrf_sp2d_ll(), and starpu_dgetrf_sp2d_rl().

◆ starpu_task_dgetrfsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	prio	TODO

Definition at line 125 of file starpu_dgetrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixNonUnit, PastixNoTrans, PastixRight, PastixUCoef, PastixUnit, PastixUpper, starpu_task_blok_dgetrf(), starpu_task_blok_dtrsmsp(), and starpu_task_cblk_dgetrfsp().

Referenced by starpu_dgetrf_sp2d_ll(), and starpu_dgetrf_sp2d_rl().

◆ starpu_dgetrf_sp1dplus_rl()

void starpu_dgetrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 183 of file starpu_dgetrf.c.

Referenced by starpu_dgetrf().

◆ starpu_dgetrf_sp1dplus_ll()

void starpu_dgetrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 251 of file starpu_dgetrf.c.

Referenced by starpu_dgetrf().

◆ starpu_dgetrf_sp2d_rl()

void starpu_dgetrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 329 of file starpu_dgetrf.c.

Referenced by starpu_dgetrf().

◆ starpu_dgetrf_sp2d_ll()

void starpu_dgetrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 428 of file starpu_dgetrf.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, PastixLUCoef, starpu_sparse_cblk_wont_use(), starpu_task_dgetrfsp(), and starpu_task_getrf_dgemmsp().

Referenced by starpu_dgetrf().

◆ starpu_dgetrf()

void starpu_dgetrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse LU factorization using StarPU runtime.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 505 of file starpu_dgetrf.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, PastixGeneral, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_dgetrf_sp1dplus_ll(), starpu_dgetrf_sp1dplus_rl(), starpu_dgetrf_sp2d_ll(), starpu_dgetrf_sp2d_rl(), starpu_sparse_matrix_getoncpu(), and starpu_sparse_matrix_init().

◆ starpu_task_getrf_sgemmsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	fcblk	TODO
[in]	blokA	TODO
[in]	prio	TODO

Definition at line 55 of file starpu_sgetrf.c.

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

Referenced by starpu_sgetrf_sp2d_ll(), and starpu_sgetrf_sp2d_rl().

◆ starpu_task_sgetrfsp()

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

TODO.

Parameters

[in]	sopalin_data	TODO
[in]	cblk	TODO
[in]	prio	TODO

Definition at line 125 of file starpu_sgetrf.c.

References solver_cblk_s::cblktype, solver_cblk_s::fblokptr, pastix_int_t, PastixLCoef, PastixNonUnit, PastixNoTrans, PastixRight, PastixUCoef, PastixUnit, PastixUpper, starpu_task_blok_sgetrf(), starpu_task_blok_strsmsp(), and starpu_task_cblk_sgetrfsp().

Referenced by starpu_sgetrf_sp2d_ll(), and starpu_sgetrf_sp2d_rl().

◆ starpu_sgetrf_sp1dplus_rl()

void starpu_sgetrf_sp1dplus_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 183 of file starpu_sgetrf.c.

Referenced by starpu_sgetrf().

◆ starpu_sgetrf_sp1dplus_ll()

void starpu_sgetrf_sp1dplus_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

Perform a sparse LU factorization with 1D kernels.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 251 of file starpu_sgetrf.c.

Referenced by starpu_sgetrf().

◆ starpu_sgetrf_sp2d_rl()

void starpu_sgetrf_sp2d_rl	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 329 of file starpu_sgetrf.c.

Referenced by starpu_sgetrf().

◆ starpu_sgetrf_sp2d_ll()

void starpu_sgetrf_sp2d_ll	(	sopalin_data_t *	sopalin_data,
		starpu_sparse_matrix_desc_t *	desc
	)

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

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 428 of file starpu_sgetrf.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, PastixLUCoef, starpu_sparse_cblk_wont_use(), starpu_task_getrf_sgemmsp(), and starpu_task_sgetrfsp().

Referenced by starpu_sgetrf().

◆ starpu_sgetrf()

void starpu_sgetrf	(	pastix_data_t *	pastix_data,
		sopalin_data_t *	sopalin_data
	)

Perform a sparse LU factorization using StarPU runtime.

The function performs the LU factorization of a sparse general matrix A. The factorization has the form

\[ A = L\times U \]

where L is a sparse lower triangular matrix, and U a sparse upper triangular with the same pattern as L^t.

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 505 of file starpu_sgetrf.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, PastixGeneral, PastixTraceNumfact, pastix_data_s::solvmatr, starpu_sparse_matrix_desc_s::solvmtx, pastix_data_s::starpu, starpu_sgetrf_sp1dplus_ll(), starpu_sgetrf_sp1dplus_rl(), starpu_sgetrf_sp2d_ll(), starpu_sgetrf_sp2d_rl(), starpu_sparse_matrix_getoncpu(), and starpu_sparse_matrix_init().

Functions

Detailed Description

Function Documentation

◆ starpu_task_getrf_zgemmsp()

◆ starpu_task_zgetrfsp()

◆ starpu_zgetrf_sp1dplus_rl()

◆ starpu_zgetrf_sp1dplus_ll()

◆ starpu_zgetrf_sp2d_rl()

◆ starpu_zgetrf_sp2d_ll()

◆ starpu_zgetrf()

◆ starpu_task_getrf_cgemmsp()

◆ starpu_task_cgetrfsp()

◆ starpu_cgetrf_sp1dplus_rl()

◆ starpu_cgetrf_sp1dplus_ll()

◆ starpu_cgetrf_sp2d_rl()

◆ starpu_cgetrf_sp2d_ll()

◆ starpu_cgetrf()

◆ starpu_task_getrf_dgemmsp()

◆ starpu_task_dgetrfsp()

◆ starpu_dgetrf_sp1dplus_rl()

◆ starpu_dgetrf_sp1dplus_ll()

◆ starpu_dgetrf_sp2d_rl()

◆ starpu_dgetrf_sp2d_ll()

◆ starpu_dgetrf()

◆ starpu_task_getrf_sgemmsp()

◆ starpu_task_sgetrfsp()

◆ starpu_sgetrf_sp1dplus_rl()

◆ starpu_sgetrf_sp1dplus_ll()

◆ starpu_sgetrf_sp2d_rl()

◆ starpu_sgetrf_sp2d_ll()

◆ starpu_sgetrf()