System matrix assembly extension module. More...

Data Types
type	sam8type
	Data type for 64-bit version of some integer arrays in sammodule::samtype. More...

Functions/Subroutines
subroutine, private	writemeqn (prefix, madof, meqn, lpu)
	Writes out the assigned equation numbers. More...

subroutine	casttoint8 (samData, ierr)
	Creates a 64-bit integer version of the SAM-arrays needed by SPR. More...

subroutine	csallocpointerarrays (samData, sysMat, matrixType, lpu, ierr, ipsw, cacheSize, needsFactorization, masterGSF)
	Initializes pointer arrays for the equation solvers. More...

subroutine	csrenumber (samData, sysMat, lpu, ierr)
	Performs nodal renumbering for the equation solver. More...

subroutine	cspreassemble (samData, sysMat, doSupelReduction, lpu, ierr, ipsw, meqnK)
	Performs pre-assembly for the given system matrix. More...

subroutine	cscreateeqpartition (is2orderedLast, samData, eqnStatus, ierr)
	Partitions the equation system into external and internal DOFs. More...

subroutine	csbeginassembly (sysMat, matName, lpu, ierr)
	Initializes the element assembly. More...

subroutine	csendassembly (sysMat, ierr)
	Finalizes the element assembly. More...

subroutine, private	csexpandelmat (elMat, expMat, mpmnpc, mmnpc, madof, iel, nedof, ierr)
	Expands an element matrix to comply with sammodule::samtype::madof. More...

subroutine	csaddelm (samData, iel, nedof, elMat, sysMat, ierr, sysRhs)
	Adds an element matrix into the system matrix. More...

subroutine	csaddem (samData, iel, nedof, elMat, sysMat, ierr, sysRhs, lpu, ipsw)
	Adds an element matrix into the system matrix. More...

subroutine	csadded (samData, iel, nedof, elMat, sysMat, ierr)
	Adds a diagonal element matrix into the system matrix. More...

subroutine	csaddev (samData, iel, nedof, eV, sysV, sysReac, ierr)
	Adds an element vector into a system vector. More...

subroutine	csaddnv (samData, inod, nodV, sysV, sysReac, ierr)
	Adds a nodal vector into a system vector. More...

subroutine	csextractelv (samData, iel, elV, sysV, nedof)
	Extracts an element vector from a system vector. More...

subroutine, private	csgetelmeq (MADOF, MEQN, MPMNPC, MMNPC, IPAR18, IEL, NEDOF, MEEN)
	Finds the matrix of element equation numbers for an element. More...

subroutine, private	cspreasssparse (samData, sparse, IERR)
	Calculates the system matrix sparsity pattern for an FE model. More...

Variables
type(sam8type), save, private	mysam
	64-bit integer version of SAM arrays More...

Detailed Description

System matrix assembly extension module.

This module contains subroutines that administer the preprocessing steps of a linear system of equations, and the assembly of element matrices and associated right-hand-side vectors into system matrices and vectors. The module defines an interface between the FEDEM solver applications and the various linear algebra packages that are being used.

Function/Subroutine Documentation

◆ casttoint8()

subroutine asmextensionmodule::casttoint8	(	type(samtype), intent(in), target	samData,
		integer, intent(out), optional	ierr
	)

Creates a 64-bit integer version of the SAM-arrays needed by SPR.

Parameters

[in]	samData	Data for managing system matrix assembly
[out]	ierr	Error flag

Author: Knut Morten Okstad

Date: 24 May 2017

Here is the caller graph for this function:

◆ csadded()

subroutine asmextensionmodule::csadded	(	type(samtype), intent(in)	samData,
		integer, intent(in)	iel,
		integer, intent(in)	nedof,
		real(dp), dimension(:), intent(in)	elMat,
		type(sysmatrixtype), intent(inout)	sysMat,
		integer, intent(out)	ierr
	)

Adds a diagonal element matrix into the system matrix.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	iel	Element index
[in]	nedof	Number of DOFs in element
[in]	elMat	The diagonal element matrix to add
	sysMat	The system matrix to add into
[out]	ierr	Error flag

Author: Knut Morten Okstad

Date: Mar 2003

Here is the caller graph for this function:

◆ csaddelm()

subroutine asmextensionmodule::csaddelm	(	type(samtype), intent(in)	samData,
		integer, intent(in)	iel,
		integer, intent(in)	nedof,
		real(dp), dimension(nedof,nedof), intent(in)	elMat,
		type(sysmatrixtype), intent(inout)	sysMat,
		integer, intent(out)	ierr,
		real(dp), dimension(:), intent(inout), optional	sysRhs
	)

Adds an element matrix into the system matrix.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	iel	Element index
[in]	nedof	Number of DOFs in element
[in]	elMat	The element matrix to add
	sysMat	The system matrix to add into
[out]	ierr	Error flag
	sysRhs	System right-hand-side vector

This subroutine does the same as asmextensionmodule::csaddem, but uses assumed-size declaration of elMat instead. It should therefore be used if the calling subroutine declares elMat as a 2D array elMat(N,N), where N > nedof, and elMat(1:nedof,1:nedof) is passed as argument. Calling asmextensionmodule::csaddem directly will not work in this case.

Knut Morten Okstad

Date: 28 Oct 2002

Here is the caller graph for this function:

◆ csaddem()

subroutine asmextensionmodule::csaddem	(	type(samtype), intent(in)	samData,
		integer, intent(in)	iel,
		integer, intent(in)	nedof,
		real(dp), dimension(:,:), intent(in)	elMat,
		type(sysmatrixtype), intent(inout)	sysMat,
		integer, intent(out)	ierr,
		real(dp), dimension(:), intent(inout), optional, target	sysRhs,
		integer, intent(in), optional	lpu,
		integer, intent(in), optional	ipsw
	)

Adds an element matrix into the system matrix.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	iel	Element index
[in]	nedof	Number of DOFs in element
[in]	elMat	The element matrix to add
	sysMat	The system matrix to add into
[out]	ierr	Error flag
	sysRhs	System right-hand-side vector
[in]	lpu	File unit number for debug print
[in]	ipsw	Print switch

Note: The argument elMat is here declared assumed-shape (KMO 21.10.02).

Author: Bjorn Haugen

Date: Sep 1998

Author: Knut Morten Okstad

Date: Mar 2003

Here is the caller graph for this function:

◆ csaddev()

subroutine asmextensionmodule::csaddev	(	type(samtype), intent(in)	samData,
		integer, intent(in)	iel,
		integer, intent(in)	nedof,
		real(dp), dimension(:), intent(in)	eV,
		real(dp), dimension(:), intent(inout)	sysV,
		real(dp), dimension(:), intent(inout)	sysReac,
		integer, intent(out)	ierr
	)

Adds an element vector into a system vector.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	iel	Element index
[in]	nedof	Number of DOFs in element
[in]	eV	The element vector to add
	sysV	The system vector to add into
	sysReac	Vector of reaction forces
[out]	ierr	Error flag

This subroutine takes into account free and dependent DOFs. Contributions from fixed or prescribed DOFs are added into sysReac.

Author: Karl Erik Thoresen

Date: Dec 1998

Author: Knut Morten Okstad

Date: Oct 2005

Here is the caller graph for this function:

◆ csaddnv()

subroutine asmextensionmodule::csaddnv	(	type(samtype), intent(in)	samData,
		integer, intent(in)	inod,
		real(dp), dimension(:), intent(in)	nodV,
		real(dp), dimension(:), intent(inout)	sysV,
		real(dp), dimension(:), intent(inout)	sysReac,
		integer, intent(out)	ierr
	)

Adds a nodal vector into a system vector.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	inod	Node index
[in]	nodV	The nodal vector to add
	sysV	The system vector to add into
	sysReac	Vector of reaction forces
[out]	ierr	Error flag

This subroutine takes into account free and dependent DOFs. Contributions from fixed or prescribed DOFs are added into sysReac.

Author: Knut Morten Okstad

Date: Oct 2005

Here is the caller graph for this function:

◆ csallocpointerarrays()

subroutine asmextensionmodule::csallocpointerarrays	(	type(samtype), intent(inout), target	samData,
		type(sysmatrixtype), intent(out)	sysMat,
		integer, intent(in)	matrixType,
		integer, intent(in)	lpu,
		integer, intent(out)	ierr,
		integer, intent(in), optional	ipsw,
		integer, intent(in), optional	cacheSize,
		logical, intent(in), optional	needsFactorization,
		type(gsfstoragetype), optional, pointer	masterGSF
	)

Initializes pointer arrays for the equation solvers.

Parameters

	samData	Data for managing system matrix assembly
[out]	sysMat	The system matrix to initialize data structure for
[in]	matrixType	Type of system matrix data structure
[in]	lpu	File unit number for res-file output
[out]	ierr	Error flag
[in]	ipsw	Print switch
[in]	cacheSize	Sparse matricx cache size
[in]	needsFactorization	If .true., this matrix will be factorized
[in]	masterGSF	GSF data structure of sibling system matrix

The symbolic assembly stage is also performed for the sparse matrix equation solvers.

Author: Bjorn Haugen

Date: Sep 1998

Author: Knut Morten Okstad

Date: Feb 2003

Here is the caller graph for this function:

◆ csbeginassembly()

subroutine asmextensionmodule::csbeginassembly	(	type(sysmatrixtype), intent(inout)	sysMat,
		character(len=*), intent(in)	matName,
		integer, intent(in)	lpu,
		integer, intent(out)	ierr
	)

Initializes the element assembly.

Parameters

	sysMat	The system matrix to initialize element assembly for
[in]	matName	Matrix name, used for storage requirement print
[in]	lpu	File unit number for res-file output
[out]	ierr	Error flag

Knut Morten Okstad

Date: 8 Sep 2004

Here is the caller graph for this function:

◆ cscreateeqpartition()

subroutine asmextensionmodule::cscreateeqpartition	(	logical, intent(in)	is2orderedLast,
		type(samtype), intent(inout)	samData,
		integer, dimension(:), intent(out)	eqnStatus,
		integer, intent(out)	ierr
	)

Partitions the equation system into external and internal DOFs.

Parameters

[in]	is2orderedLast	If .true., the status-2 DOFs are orered last
	samData	Data for managing system matrix assembly
[out]	eqnStatus	Status flags for the equations
[out]	ierr	Error flag

Knut Morten Okstad

Date: 16 Jan 2003

Here is the caller graph for this function:

◆ csendassembly()

subroutine asmextensionmodule::csendassembly	(	type(sysmatrixtype), intent(inout)	sysMat,
		integer, intent(out)	ierr
	)

Finalizes the element assembly.

Parameters

	sysMat	The system matrix to finalize element assembly for
[out]	ierr	Error flag

Knut Morten Okstad

Date: 8 Sep 2004

Here is the caller graph for this function:

◆ csexpandelmat()

subroutine, private asmextensionmodule::csexpandelmat	(	real(dp), dimension(:,:), intent(in), target	elMat,
		real(dp), dimension(:,:), pointer	expMat,
		integer, dimension(:), intent(in)	mpmnpc,
		integer, dimension(:), intent(in)	mmnpc,
		integer, dimension(:), intent(in)	madof,
		integer, intent(in)	iel,
		integer, intent(in)	nedof,
		integer, intent(out)	ierr
	)

private

Expands an element matrix to comply with sammodule::samtype::madof.

Parameters

[in]	elMat	Element stiffness matrix as returned from element routine
[out]	expMat	Expanded elemetn matrix
[in]	mpmnpc	Matrix of pointers to MNPCs
[in]	mmnpc	Matrix of nodal point correspondances
[in]	madof	Matrix of accumulated DOFs
[in]	iel	Element index
[in]	nedof	Number of rows and columns in elMat
[out]	ierr	Error flag

The element matrix elMat is expanded such that it matches the number of DOFs per node as defined by the sam array MADOF.

Knut Morten Okstad

Date: 2 Jun 2005

Here is the caller graph for this function:

◆ csextractelv()

subroutine asmextensionmodule::csextractelv	(	type(samtype), intent(in)	samData,
		integer, intent(in)	iel,
		real(dp), dimension(:), intent(out)	elV,
		real(dp), dimension(:), intent(in)	sysV,
		integer, intent(out), optional	nedof
	)

Extracts an element vector from a system vector.

Parameters

[in]	samData	Data for managing system matrix assembly
[in]	iel	Element index
[out]	elV	Extracted element vector.
[in]	sysV	The system vector to extract from
[out]	nedof	Total number of DOFs in element

Author: Karl Erik Thoresen

Date: Dec 1998

Here is the caller graph for this function:

◆ csgetelmeq()

subroutine, private asmextensionmodule::csgetelmeq	(	integer, dimension(:), intent(in)	MADOF,
		integer, dimension(:), intent(in)	MEQN,
		integer, dimension(:), intent(in)	MPMNPC,
		integer, dimension(:), intent(in)	MMNPC,
		integer, intent(in)	IPAR18,
		integer, intent(in)	IEL,
		integer, intent(out)	NEDOF,
		integer, dimension(:), intent(out), optional	MEEN
	)

private

Finds the matrix of element equation numbers for an element.

Parameters

[in]	MADOF	Matrix of accumulated DOFs
[in]	MEQN	Matrix of equation numbers
[in]	MPMNPC	Matrix of pointers to MNPCs
[in]	MMNPC	Matrix of nodal point correspondances
[in]	IPAR18	Value of mpar(18)
[in]	IEL	Element index
[out]	NEDOF	Total number of DOFs in element
[out]	MEEN	Matrix of equation numbers associated with the element

Author: Knut Morten Okstad

Date: 12 Feb 2016

Here is the caller graph for this function:

◆ cspreassemble()

subroutine asmextensionmodule::cspreassemble	(	type(samtype), intent(inout)	samData,
		type(sysmatrixtype), intent(inout)	sysMat,
		logical, intent(in)	doSupelReduction,
		integer, intent(in)	lpu,
		integer, intent(out)	ierr,
		integer, intent(in), optional	ipsw,
		integer, dimension(:), intent(in), optional	meqnK
	)

Performs pre-assembly for the given system matrix.

Parameters

	samData	Data for managing system matrix assembly
	sysMat	The system matrix to perform per-assembly for
[in]	doSupelReduction	If .true., we are doing superelement reduction
[in]	lpu	File unit number for res-file output
[out]	ierr	Error flag
[in]	ipsw	Print switch
[in]	meqnK	Matrix of equation numbers for system stiffness matrix

Author: Bjorn Haugen

Date: Sep 1998

Author: Knut Morten Okstad

Date: Jan 2003

Here is the caller graph for this function:

◆ cspreasssparse()

subroutine, private asmextensionmodule::cspreasssparse	(	type(samtype), intent(in)	samData,
		type(pardisostoragetype), intent(inout)	sparse,
		integer, intent(out)	IERR
	)

private

Calculates the system matrix sparsity pattern for an FE model.

Parameters

[in]	samData	Data for managing system matrix assembly
	sparse	sparse Sparse data structure for the Pardiso equation solver
[out]	ierr	Error flag

Author: Knut Morten Okstad

Date: 12 Feb 2016

Here is the caller graph for this function:

◆ csrenumber()

subroutine asmextensionmodule::csrenumber	(	type(samtype), intent(inout)	samData,
		type(sysmatrixtype), intent(inout)	sysMat,
		integer, intent(in)	lpu,
		integer, intent(out)	ierr
	)

Performs nodal renumbering for the equation solver.

Parameters

	samData	Data for managing system matrix assembly
	sysMat	The system matrix to perform renumbering for
[in]	lpu	File unit number for res-file output
[out]	ierr	Error flag

Author: Bjorn Haugen

Date: Sep 1998

Author: Knut Morten Okstad

Date: Jan 2003

Here is the caller graph for this function:

◆ writemeqn()

subroutine, private asmextensionmodule::writemeqn	(	character(len=*), intent(in)	prefix,
		integer, dimension(:), intent(in)	madof,
		integer, dimension(:), intent(in)	meqn,
		integer, intent(in)	lpu
	)

private

Writes out the assigned equation numbers.

Parameters

[in]	prefix	Message prefix
[in]	madof	Matrix of accumulated DOFs
[in]	meqn	Matrix of equation numbers
[in]	lpu	File unit number for res-file output

Author: Knut Morten Okstad

Date: 28 May 2018

Here is the caller graph for this function:

Variable Documentation

◆ mysam

type(sam8type), save, private asmextensionmodule::mysam

private

64-bit integer version of SAM arrays

Data Types

Functions/Subroutines

Variables

Detailed Description

Function/Subroutine Documentation

◆ casttoint8()

◆ csadded()

◆ csaddelm()

◆ csaddem()

◆ csaddev()

◆ csaddnv()

◆ csallocpointerarrays()

◆ csbeginassembly()

◆ cscreateeqpartition()

◆ csendassembly()

◆ csexpandelmat()

◆ csextractelv()

◆ csgetelmeq()

◆ cspreassemble()

◆ cspreasssparse()

◆ csrenumber()

◆ writemeqn()

Variable Documentation

◆ mysam