lapl_operator_oct_m Module Reference

Computes \(y = -\Delta(x) \) and \(y = -\Delta(x) + shift x \) , suitable as an operator callback for iterative solvers (CG, QMR, etc.). More...

Detailed Description

Computes \(y = -\Delta(x) \) and \(y = -\Delta(x) + shift x \) , suitable as an operator callback for iterative solvers (CG, QMR, etc.).

Functions/Subroutines
subroutine, public	zlaplacian_op (x, hx, userdata)
	Computes the negative Laplacian operator action: \( hx = -\nabla^2 x \). More...
subroutine, public	zshifted_laplacian_op (x, lx, userdata)
	Computes the shifted Laplacian operator: \( lx = -\nabla^2 x + \sigma x \). More...
subroutine, public	dlaplacian_op (x, hx, userdata)
	Computes the negative Laplacian operator action: \( hx = -\nabla^2 x \). More...
subroutine, public	dshifted_laplacian_op (x, lx, userdata)
	Computes the shifted Laplacian operator: \( lx = -\nabla^2 x + \sigma x \). More...

Function/Subroutine Documentation

zlaplacian_op()

subroutine, public lapl_operator_oct_m::zlaplacian_op	(	complex(real64), dimension(:), intent(in), contiguous	x,
		complex(real64), dimension(:), intent(out), contiguous	hx,
		type(c_ptr), dimension(:), intent(in)	userdata
	)

Computes the negative Laplacian operator action: \( hx = -\nabla^2 x \).

This subroutine is designed as an operator callback for iterative solvers (conjugate gradients, QMR, etc.) and conforms to the zoperator_i interface defined in solvers_oct_m.

Parameters

[in]	x	Input vector (size: dermeshnp)
[out]	hx	Output vector: \( hx = -\nabla^2 x \) (size: dermeshnp)
[in]	userdata	Array of C pointers with FIXED indices for passing data: userdata(1): c_ptr to derivatives_t derivative with mesh information

Note

The userdata must be passed as a TARGET variable (not an expression) since c_loc requires an addressable memory location.

Definition at line 225 of file lapl_operator.F90.

zshifted_laplacian_op()

subroutine, public lapl_operator_oct_m::zshifted_laplacian_op	(	complex(real64), dimension(:), intent(in), contiguous	x,
		complex(real64), dimension(:), intent(out), contiguous	lx,
		type(c_ptr), dimension(:), intent(in)	userdata
	)

Computes the shifted Laplacian operator: \( lx = -\nabla^2 x + \sigma x \).

This subroutine implements the operator \((-\Delta + \sigma)\) where \(\sigma\) is a scalar shift parameter. It is designed as an operator callback for iterative solvers and conforms to the zoperator_i interface defined in solvers_oct_m.

Parameters

[in]	x	Input vector (size: dermeshnp)
[out]	lx	Output vector: \( lx = -\nabla^2 x + \sigma \cdot x \) (size: dermeshnp)
[in]	userdata	Array of C pointers with FIXED indices for passing data: userdata(1): c_ptr to derivatives_t derivative with mesh information userdata(2): c_ptr to real(real64) shift parameter \(\sigma\) userdata(3): c_ptr to integer prefactor which scales the entire operator

Note

The userdata must be passed as a TARGET variable (not an expression) since c_loc requires an addressable memory location.

Parameters

[in]	x	Current guess for preconditioned residual
[out]	lx	The LHS product (in CG: Ax, or (\Delta - q_0^2) f_out)

Definition at line 266 of file lapl_operator.F90.

dlaplacian_op()

subroutine, public lapl_operator_oct_m::dlaplacian_op	(	real(real64), dimension(:), intent(in), contiguous	x,
		real(real64), dimension(:), intent(out), contiguous	hx,
		type(c_ptr), dimension(:), intent(in)	userdata
	)

Computes the negative Laplacian operator action: \( hx = -\nabla^2 x \).

This subroutine is designed as an operator callback for iterative solvers (conjugate gradients, QMR, etc.) and conforms to the doperator_i interface defined in solvers_oct_m.

Parameters

[in]	x	Input vector (size: dermeshnp)
[out]	hx	Output vector: \( hx = -\nabla^2 x \) (size: dermeshnp)
[in]	userdata	Array of C pointers with FIXED indices for passing data: userdata(1): c_ptr to derivatives_t derivative with mesh information

Note

The userdata must be passed as a TARGET variable (not an expression) since c_loc requires an addressable memory location.

Definition at line 385 of file lapl_operator.F90.

dshifted_laplacian_op()

subroutine, public lapl_operator_oct_m::dshifted_laplacian_op	(	real(real64), dimension(:), intent(in), contiguous	x,
		real(real64), dimension(:), intent(out), contiguous	lx,
		type(c_ptr), dimension(:), intent(in)	userdata
	)

Computes the shifted Laplacian operator: \( lx = -\nabla^2 x + \sigma x \).

This subroutine implements the operator \((-\Delta + \sigma)\) where \(\sigma\) is a scalar shift parameter. It is designed as an operator callback for iterative solvers and conforms to the doperator_i interface defined in solvers_oct_m.

Parameters

[in]	x	Input vector (size: dermeshnp)
[out]	lx	Output vector: \( lx = -\nabla^2 x + \sigma \cdot x \) (size: dermeshnp)
[in]	userdata	Array of C pointers with FIXED indices for passing data: userdata(1): c_ptr to derivatives_t derivative with mesh information userdata(2): c_ptr to real(real64) shift parameter \(\sigma\) userdata(3): c_ptr to integer prefactor which scales the entire operator

Note

The userdata must be passed as a TARGET variable (not an expression) since c_loc requires an addressable memory location.

Parameters

[in]	x	Current guess for preconditioned residual
[out]	lx	The LHS product (in CG: Ax, or (\Delta - q_0^2) f_out)

Definition at line 426 of file lapl_operator.F90.