Smoothers.src File Reference

Data Types
module	smoothers

Functions/Subroutines
real(kind=dp) function	mgnorm (n, x)
subroutine	cg (n, A, M, x, b, r, Rounds)
subroutine	uzawamv (A, x, b)
subroutine	solvelinsys (N, LDa, A, x)
real(kind=dp) function	mgdot (n, x, y)
complex(kind=dp) function	mgcdot (n, x, y)
subroutine	mgmv (A, x, b, Update)
subroutine	mgcmv (A, x, b, Update)
subroutine	jacobi (n, A, M, x, b, r, Rounds)
subroutine	smoothedjacobi (n, A, M, x, b, r, w, Rounds)
subroutine	cjacobi (n, A, M, rx, rb, rr, Rounds)
subroutine	gs (n, A, M, x, b, r, Rounds)
subroutine	richards (n, A, M, x, b, r, Rounds)
subroutine	bgs (n, A, M, x, b, r, DOFs, Rounds)
subroutine	smoothedgs (n, A, M, x, b, r, w, Rounds)
subroutine	sgs (n, A, M, x, b, r, Rounds)
subroutine	internalsgs (n, A, M, x, b, r, Rounds)
subroutine	bsgs (n, A, M, x, b, r, DOFs, Rounds)
subroutine	smoothedsgs (n, A, M, x, b, r, w, Rounds)
subroutine	csgs (n, A, M, rx, rb, rr, Rounds)
subroutine	postsgs (n, A, M, x, b, r, f, Rounds)
subroutine	direct1dsmoother (n, A, M, x, b, r, f, Rounds)
subroutine	ccg (n, A, M, rx, rb, rr, Rounds)
subroutine	uzawa (n, A, M, x, b, r, Rounds)
subroutine	uzawapcond (A, b)
subroutine	bicguzawa (n, A, M, x, b, r, Rounds, reps)
subroutine	bicg (n, A, M, x, b, r, Rounds)
subroutine	vanka (n, A, M, x, b, r, Rounds)
subroutine	testgs (n, A, M, x, b, r, Rounds)

Function/Subroutine Documentation

subroutine mgsmooth::bgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	DOFs,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::bicg	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

References crsmatrix::crs_lusolve(), mgdot(), and mgmv().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::bicguzawa	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds,
		real(kind=dp)	reps
	)

References mgdot(), uzawamv(), and uzawapcond().

Referenced by uzawa().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::bsgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	DOFs,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::ccg	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	rx,
		real(kind=dp), dimension(:)	rb,
		real(kind=dp), dimension(:)	rr,
		integer	Rounds
	)

Complex valued conjugate gradient as a smoother.

References crsmatrix::crs_complexlusolve(), mgcdot(), and mgcmv().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::cg	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Conjugate gradient as a smoother.

References crsmatrix::crs_lusolve(), mgdot(), and mgmv().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::cjacobi	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	rx,
		real(kind=dp), dimension(:)	rb,
		real(kind=dp), dimension(:)	rr,
		integer	Rounds
	)

References mgcmv().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::csgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	rx,
		real(kind=dp), dimension(:)	rb,
		real(kind=dp), dimension(:)	rr,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::direct1dsmoother	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer, dimension(:), pointer	f,
		integer	Rounds
	)

For some cases a smoother that only works with the local strong connections might be ideal. Given the clustering "f" just picks the entries i and j such that f(i)=f(j) and use the reduced matrix in a direct solver to smooth the system. Note that the method guarantees that the linear system actually consists of a number of local problems that are fairly small in size.

References meshutils::detectextrudedstructure(), directsolve::directsolver(), messages::fatal(), messages::info(), mgmv(), and solver().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::gs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::internalsgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Internal symmetric-gauss-seidel for parallel computations.

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::jacobi	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

References mgmv().

Here is the call graph for this function:

complex(kind=dp) function mgsmooth::mgcdot	(	integer	n,
		complex(kind=dp), dimension(:)	x,
		complex(kind=dp), dimension(:)	y
	)

Referenced by ccg().

Here is the caller graph for this function:

subroutine mgsmooth::mgcmv	(	type(matrix_t), pointer	A,
		complex(kind=dp), dimension(:)	x,
		complex(kind=dp), dimension(:)	b,
		logical, optional	Update
	)

References crsmatrix::crs_complexmatrixvectormultiply().

Here is the call graph for this function:

real(kind=dp) function mgsmooth::mgdot	(	integer	n,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	y
	)

References parallelutils::paralleldot().

Referenced by bicg(), paralleleigensolve::bicgpareigen(), bicguzawa(), cg(), and paralleleigensolve::cgpareigen().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::mgmv	(	type(matrix_t), pointer	A,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		logical, optional	Update
	)

References crsmatrix::crs_matrixvectormultiply(), and parallelutils::parallelmatrixvector().

Here is the call graph for this function:

real(kind=dp) function mgsmooth::mgnorm	(	integer	n,
		real(kind=dp), dimension(:)	x
	)

References parallelutils::parallelnorm().

Here is the call graph for this function:

subroutine mgsmooth::postsgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer, dimension(:), pointer	f,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::richards	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Richards iteration with preconditioning by lumped mass matrix.

subroutine mgsmooth::sgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

subroutine mgsmooth::smoothedgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		real(kind=dp)	w,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::smoothedjacobi	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		real(kind=dp)	w,
		integer	Rounds
	)

References mgmv().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::smoothedsgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		real(kind=dp)	w,
		integer	Rounds
	)

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::solvelinsys	(	integer	N,
		integer	LDa,
		double precision, dimension(lda,*)	A,
		double precision, dimension(n)	x
	)

References messages::info().

Referenced by defutils::defaultdirichletbcs(), and vanka().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::testgs	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Gauss-Seidel smoother with plenty of comments for testing purposes.

Referenced by smoothers::mgsmooth().

Here is the caller graph for this function:

subroutine mgsmooth::uzawa	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

Uzawa algorithm as a smoother.

References bicguzawa(), mgmv(), and lists::variableget().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine mgsmooth::uzawamv	(	type(matrix_t)	A,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b
	)

Referenced by bicguzawa().

Here is the caller graph for this function:

subroutine mgsmooth::uzawapcond	(	type(matrix_t)	A,
		real(kind=dp), dimension(:)	b
	)

Referenced by bicguzawa().

Here is the caller graph for this function:

subroutine mgsmooth::vanka	(	integer	n,
		type(matrix_t), pointer	A,
		type(matrix_t), pointer	M,
		real(kind=dp), dimension(:)	x,
		real(kind=dp), dimension(:)	b,
		real(kind=dp), dimension(:)	r,
		integer	Rounds
	)

References crsmatrix::crs_getmatrixelement(), generalutils::getvarname(), solvelinsys(), and lists::variableget().

Referenced by smoothers::mgsmooth().

Here is the call graph for this function:

Here is the caller graph for this function: