fem_composite Module Reference

Detailed Description

Classes and infrastructure for composite FE representations.

Authors

Chris Hansen

Date

April 2014

Data Types
interface	ml_fe_base_pop
	Needs docs. More...
interface	ml_fe_base_push
	Needs docs. More...
type	oft_fem_comp_ptr
	Composite FE type pointer container. More...
type	oft_fem_comp_type
	Composite FE type. More...
type	oft_ml_fe_comp_vecspace
	Needs docs. More...
type	oft_ml_fem_comp_type
	Multi-level composite FE type. More...

Functions/Subroutines
subroutine	fem_comp_delete (self)
	Needs docs.
subroutine	fem_mat_add_local (self, mat, mloc, iloc, tlocks)
	Add local contributions to full matrix in a thread-safe way.
subroutine	fem_mat_create (self, new, mask)
	Needs docs.
subroutine	fem_mat_destroy_local (self, mloc)
	Needs docs.
subroutine	fem_mat_setup_local (self, mloc, mask)
	Needs docs.
subroutine	fem_mat_zero_local (self, mloc)
	Needs docs.
subroutine	fem_mat_zero_local_rows (self, mloc, flag, irow)
	Needs docs.
subroutine	fem_vec_create (self, new, cache, native)
	Create weight vector for FE representation.
subroutine	fem_vec_load (self, source, filename, path, err_flag)
	Load a Lagrange scalar field from a HDF5 restart file.
subroutine	fem_vec_save (self, source, filename, path, append)
	Save a Lagrange scalar field to a HDF5 restart file.
subroutine	ml_fe_vecspace_create (self, new, level, cache, native)
	Needs docs.
subroutine	ml_fe_vecspace_inject (self, afine, acors)
	Interpolate a coarse level Lagrange scalar field to the next finest level.
subroutine	ml_fe_vecspace_interp (self, acors, afine)
	Interpolate a coarse level Lagrange scalar field to the next finest level.
subroutine	ml_fem_build_interp (self, minlev)
	Needs docs.
subroutine	ml_fem_comp_delete (self)
	Needs docs.
subroutine	ml_fem_set_level (self, level, propogate)
	Set the current level for a ML Compsite-FE structure.
subroutine	ml_fem_setup (self)
	Needs docs.
subroutine	ml_fem_vec_create (self, new, level, cache, native)
	Create weight vector for FE representation.

Function/Subroutine Documentation

fem_comp_delete()

subroutine fem_comp_delete

(

class(oft_fem_comp_type), intent(inout)

self

)

Needs docs.

fem_mat_add_local()

subroutine fem_mat_add_local	(	class(oft_fem_comp_type), intent(inout)	self,
		class(oft_matrix), intent(inout)	mat,
		type(oft_local_mat), dimension(:,:), intent(in)	mloc,
		type(oft_1d_int), dimension(:), intent(in)	iloc,
		integer(kind=omp_lock_kind), dimension(:), intent(inout)	tlocks
	)

Add local contributions to full matrix in a thread-safe way.

Parameters

[in,out]	mat	Full matrix
[in]	mloc	Local matrix
[in]	iloc	Local FE entries
[in,out]	tlocks	OpenMP row thread locks

fem_mat_create()

subroutine fem_mat_create	(	class(oft_fem_comp_type), intent(inout)	self,
		class(oft_matrix), intent(out), pointer	new,
		integer(i4), dimension(:,:), intent(in), optional	mask
	)

Needs docs.

fem_mat_destroy_local()

subroutine fem_mat_destroy_local	(	class(oft_fem_comp_type), intent(inout)	self,
		type(oft_local_mat), dimension(:,:), intent(inout)	mloc
	)

Needs docs.

fem_mat_setup_local()

subroutine fem_mat_setup_local	(	class(oft_fem_comp_type), intent(inout)	self,
		type(oft_local_mat), dimension(:,:), intent(inout)	mloc,
		integer(i4), dimension(:,:), intent(in), optional	mask
	)

Needs docs.

fem_mat_zero_local()

subroutine fem_mat_zero_local	(	class(oft_fem_comp_type), intent(inout)	self,
		type(oft_local_mat), dimension(:,:), intent(inout)	mloc
	)

Needs docs.

fem_mat_zero_local_rows()

subroutine fem_mat_zero_local_rows	(	class(oft_fem_comp_type), intent(inout)	self,
		type(oft_local_mat), dimension(:,:), intent(inout)	mloc,
		logical, dimension(:), intent(in)	flag,
		integer(i4), intent(in)	irow
	)

Needs docs.

fem_vec_create()

subroutine fem_vec_create	(	class(oft_fem_comp_type), intent(inout)	self,
		class(oft_vector), intent(out), pointer	new,
		logical, intent(in), optional	cache,
		logical, intent(in), optional	native
	)

Create weight vector for FE representation.

Parameters

[out]	new	Vector to create
[in]	cache	Allow caching (optional)
[in]	native	Force native representation (optional)

fem_vec_load()

subroutine fem_vec_load	(	class(oft_fem_comp_type), intent(inout)	self,
		class(oft_vector), intent(inout), target	source,
		character(len=*), intent(in)	filename,
		character(len=*), intent(in)	path,
		integer(i4), intent(out), optional	err_flag
	)

Load a Lagrange scalar field from a HDF5 restart file.

Parameters

[in,out]	source	Destination field
[in]	filename	Name of source file
[in]	path	ield path in file
[out]	err_flag	Error flag

fem_vec_save()

subroutine fem_vec_save	(	class(oft_fem_comp_type), intent(inout)	self,
		class(oft_vector), intent(inout), target	source,
		character(len=*), intent(in)	filename,
		character(len=*), intent(in)	path,
		logical, intent(in), optional	append
	)

Save a Lagrange scalar field to a HDF5 restart file.

Parameters

[in,out]	source	Source field
[in]	filename	Name of destination file
[in]	path	Field label in file
[in]	append	Append to file instead of creating?

ml_fe_vecspace_create()

subroutine ml_fe_vecspace_create	(	class(oft_ml_fe_comp_vecspace), intent(inout)	self,
		class(oft_vector), intent(out), pointer	new,
		integer(i4), intent(in), optional	level,
		logical, intent(in), optional	cache,
		logical, intent(in), optional	native
	)

Needs docs.

ml_fe_vecspace_inject()

subroutine ml_fe_vecspace_inject	(	class(oft_ml_fe_comp_vecspace), intent(inout)	self,
		class(oft_vector), intent(inout)	afine,
		class(oft_vector), intent(inout)	acors
	)

Interpolate a coarse level Lagrange scalar field to the next finest level.

Note

The global Lagrange level in incremented by one in this subroutine

Parameters

[in,out]	afine	Fine vector from interpolation
[in,out]	acors	Vector to interpolate

ml_fe_vecspace_interp()

subroutine ml_fe_vecspace_interp	(	class(oft_ml_fe_comp_vecspace), intent(inout)	self,
		class(oft_vector), intent(inout)	acors,
		class(oft_vector), intent(inout)	afine
	)

Interpolate a coarse level Lagrange scalar field to the next finest level.

Note

The global Lagrange level in incremented by one in this subroutine

Parameters

[in,out]	acors	Vector to interpolate
[in,out]	afine	Fine vector from interpolation

ml_fem_build_interp()

subroutine ml_fem_build_interp	(	class(oft_ml_fem_comp_type), intent(inout)	self,
		integer(i4), intent(in), optional	minlev
	)

Needs docs.

ml_fem_comp_delete()

subroutine ml_fem_comp_delete

(

class(oft_ml_fem_comp_type), intent(inout)

self

)

Needs docs.

ml_fem_set_level()

subroutine ml_fem_set_level	(	class(oft_ml_fem_comp_type), intent(inout)	self,
		integer(i4), intent(in)	level,
		logical, intent(in), optional	propogate
	)

Set the current level for a ML Compsite-FE structure.

Parameters

[in]

level

Desired level

ml_fem_setup()

subroutine ml_fem_setup

(

class(oft_ml_fem_comp_type), intent(inout)

self

)

Needs docs.

ml_fem_vec_create()

subroutine ml_fem_vec_create	(	class(oft_ml_fem_comp_type), intent(inout)	self,
		class(oft_vector), intent(out), pointer	new,
		integer(i4), intent(in), optional	level,
		logical, intent(in), optional	cache,
		logical, intent(in), optional	native
	)

Create weight vector for FE representation.

Parameters

[out]	new	Vector to create
[in]	level	FE level for init (optional)
[in]	cache	Allow caching (optional)
[in]	native	Force native representation (optional)