wrap Trilinos solver functions More...

#include <stdio.h>
#include <vector>
#include <iostream>
#include "mpi.h"
#include <time.h>
#include <numeric>
#include "Epetra_MpiComm.h"
#include "Epetra_Map.h"
#include "Epetra_FEVbrMatrix.h"
#include "Epetra_FEVector.h"
#include "Epetra_FECrsGraph.h"
#include "Epetra_FECrsMatrix.h"
#include "Epetra_Import.h"
#include "AztecOO.h"
#include "AztecOO_StatusTestResNorm.h"
#include "ml_include.h"
#include "ml_MultiLevelPreconditioner.h"
#include "Ifpack.h"
#include "Ifpack_ConfigDefs.h"
#include "Ifpack_AdditiveSchwarz.h"
#include "Ifpack_ILUT.h"
#include "Ifpack_IC.h"
#include "Ifpack_ICT.h"

Include dependency graph for trilinos_linear_solver.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	Trilinos
	Initialize all Epetra types we need separate from Fortran. More...

class	TrilinosMatVec
	This class implements the pure virtual class Epetra_Operator for the AztecOO iterative solve which only uses the Apply() method to compute the matrix vector product. More...

Macros
#define	TRILINOS_CG_SOLVER 798

#define	TRILINOS_GMRES_SOLVER 797

#define	TRILINOS_BICGSTAB_SOLVER 795

#define	NO_PRECONDITIONER 700

#define	TRILINOS_DIAGONAL_PRECONDITIONER 702

#define	TRILINOS_BLOCK_JACOBI_PRECONDITIONER 703

#define	TRILINOS_ILU_PRECONDITIONER 704

#define	TRILINOS_ILUT_PRECONDITIONER 705

#define	TRILINOS_IC_PRECONDITIONER 706

#define	TRILINOS_ICT_PRECONDITIONER 707

#define	TRILINOS_ML_PRECONDITIONER 708

Functions
void	trilinos_lhs_create_ (int &numGlobalNodes, int &numLocalNodes, int &numGhostAndLocalNodes, int &nnz, const int ltgSorted, const int ltgUnsorted, const int rowPtr, const int colInd, int &dof, int &cpp_index, int &proc_id)
	Give function definitions which will be called through fortran. More...

void	trilinos_bc_create_ (const double *v, bool &isCoupledBC)

void	trilinos_doassem_ (int &numNodesPerElement, const int eqN, const double lK, double *lR)

void	trilinos_global_solve_ (const double Val, const double RHS, double x, const double dirW, double &resNorm, double &initNorm, int &numIters, double &solverTime, double &dB, bool &converged, int &lsType, double &relTol, int &maxIters, int &kspace, int &precondType)

void	trilinos_solve_ (double x, const double dirW, double &resNorm, double &initNorm, int &numIters, double &solverTime, double &dB, bool &converged, int &lsType, double &relTol, int &maxIters, int &kspace, int &precondType, bool &isFassem)

void	trilinos_lhs_free_ ()

void	setPreconditioner (int precondType, AztecOO &Solver)

void	setMLPrec (AztecOO &Solver)

void	setIFPACKPrec (AztecOO &Solver)

void	checkDiagonalIsZero ()

void	constructJacobiScaling (const double *dirW, Epetra_Vector &diagonal)

void	printMatrixToFile ()

void	printRHSToFile ()

void	printSolutionToFile ()

Detailed Description

wrap Trilinos solver functions

Copyright (c) Stanford University, The Regents of the University of California, and others.

See Copyright-SimVascular.txt for additional details.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Function Documentation

◆ checkDiagonalIsZero()

void checkDiagonalIsZero ( )

This routine is to be used with preconditioners such as ILUT which require 1s on the diagonal

◆ constructJacobiScaling()

void constructJacobiScaling	(	const double *	dirW,
		Epetra_Vector &	diagonal
	)

To be called within solve-output diagonal from here

Parameters

dirW	pass in array with Dirichlet boundary face nodes marked \paramdiagonal diagonal scaling vector need to output to multiply solution by

◆ printMatrixToFile()

void printMatrixToFile ( )

for debugging purposes here are routines to print the matrix and RHS vector

◆ setIFPACKPrec()

void setIFPACKPrec ( AztecOO & Solver )

pass in IC, ICT pass in string for which to turn on right now set to IC

◆ setMLPrec()

void setMLPrec ( AztecOO & Solver )

Tune parameters for htis and IFPACK Ref: https://trilinos.org/oldsite/packages/ml/mlguide5.pdf

◆ trilinos_bc_create_()

void trilinos_bc_create_	(	const double *	v,
		bool &	isCoupledBC
	)

Parameters

v	coeff in the scalar product
isCoupledBC	determines if coupled resistance BC is turned on
v	coupled boundary vector
isCoupledBC	determines if coupled resistance BC is turned on

◆ trilinos_doassem_()

void trilinos_doassem_	(	int &	numNodesPerElement,
		const int *	eqN,
		const double *	lK,
		double *	lR
	)

This function assembles the dense element block stiffness matrix and element block force vector into the global block stiffness K and global block force vector F. Happens on a single element and will be called repeatedly as we loop over all the elements.

Parameters

numNodesPerElement	number of nodes (d) for element e
eqN	converts local element node number to local proc equation number-size is numNodesPerElement
lK	element local stiffness of size (dof*dof, numNodesPerElement, numNodesPerElement)
lR	element local force vector of size (dof, numNodesPerElement)

◆ trilinos_global_solve_()

void trilinos_global_solve_	(	const double *	Val,
		const double *	RHS,
		double *	x,
		const double *	dirW,
		double &	resNorm,
		double &	initNorm,
		int &	numIters,
		double &	solverTime,
		double &	dB,
		bool &	converged,
		int &	lsType,
		double &	relTol,
		int &	maxIters,
		int &	kspace,
		int &	precondType
	)

Tthis function creates the LHS structure topology and RHS vector based on the block map using the global stiffness matrix and global force vector assuming the assembly part has been done in svFSI

Parameters

Val	nonzero values of assembled global stiffness
RHS	global force vector
x	solution vector
dirW	gives information about the Dirichlet BC
resNorm	norm of solution
initNorm	initial norm of RHS
numIters	linear solver number of iterations
solverTime	time in linear solver
dB	log ratio for output
converged	neither or not converged in the max number of iterations
lsType	defines type of linear solver to use
relTol	default relative tolerance or as set in input file
maxIters	default max number of iterations for gmres per restart
kspace	specific for gmres dim of the stored Krylov space vectors
precondType	defines type of preconditioner to use

◆ trilinos_lhs_create_()

void trilinos_lhs_create_	(	int &	numGlobalNodes,
		int &	numLocalNodes,
		int &	numGhostAndLocalNodes,
		int &	nnz,
		const int *	ltgSorted,
		const int *	ltgUnsorted,
		const int *	rowPtr,
		const int *	colInd,
		int &	Dof,
		int &	cpp_index,
		int &	proc_id
	)

Give function definitions which will be called through fortran.

make the graph global and optimize storage on it this function creates the LHS structure topology and RHS vector based on the block map

Parameters

numGlobalNodes	total/global number of nodes each node can have dof for the spatial dim
numLocalNodes	number of nodes owned by this proc in block coordiantes
numGhostAndLocalNodes	number of nodes owned and shared by this processor includes ghost nodes
nnz	number of nonzeros in LHS matrix of calling proc
ltgSorted	integer pointer of size numLocalNodes returns global node number of local node a to give blockrow
ltgUnsorted	unsorted/not-reordered version
rowPtr	CSR row ptr of size numLocalNodes + 1 to block rows
colInd	CSR column indices ptr (size nnz points) to block rows
Dof	size of each block element to give dim of each block

◆ trilinos_lhs_free_()

void trilinos_lhs_free_ ( )

free the memory of the global data structures

◆ trilinos_solve_()

void trilinos_solve_	(	double *	x,
		const double *	dirW,
		double &	resNorm,
		double &	initNorm,
		int &	numIters,
		double &	solverTime,
		double &	dB,
		bool &	converged,
		int &	lsType,
		double &	relTol,
		int &	maxIters,
		int &	kspace,
		int &	precondType,
		bool &	isFassem
	)

This function uses the established data structure to solve the block linear system and passes the solution vector back to fortran with the ghost nodes filled in

Parameters

x	solution vector
dirW	gives information about Dirichlet BC
resNorm	norm of the final residual
initNorm	norm of initial resiual x_init = 0 \|\|b\|\|
numIters	number of iterations in the linear solver
solverTime	time in the linear solver
dB	log ratio for output
converged	can pass in solver and preconditioner type too
lsType	defines type of linear solver to use
relTol	default relative tolerance or as set in input file
maxIters	default max number of iterations for gmres per restart
kspace	specific for gmres dim of the stored Krylov space vectors
precondType	defines type of preconditioner to use
isFassem	determines if F is already assembled at ghost nodes

Classes

Macros

Functions

Detailed Description

Function Documentation

◆ checkDiagonalIsZero()

◆ constructJacobiScaling()

◆ printMatrixToFile()

◆ setIFPACKPrec()

◆ setMLPrec()

◆ trilinos_bc_create_()

◆ trilinos_doassem_()

◆ trilinos_global_solve_()

◆ trilinos_lhs_create_()

◆ trilinos_lhs_free_()

◆ trilinos_solve_()