A global linear system. More...

#include <GlobalLinSysStaticCond.h>

Inheritance diagram for Nektar::MultiRegions::GlobalLinSysStaticCond:

Collaboration diagram for Nektar::MultiRegions::GlobalLinSysStaticCond:

Public Member Functions
	GlobalLinSysStaticCond (const GlobalLinSysKey &mkey, const boost::weak_ptr< ExpList > &pExpList, const boost::shared_ptr< AssemblyMap > &locToGloMap)
	Constructor for full direct matrix solve. More...

virtual	~GlobalLinSysStaticCond ()

Public Member Functions inherited from Nektar::MultiRegions::GlobalLinSys
	GlobalLinSys (const GlobalLinSysKey &pKey, const boost::weak_ptr< ExpList > &pExpList, const boost::shared_ptr< AssemblyMap > &pLocToGloMap)
	Constructor for full direct matrix solve. More...

virtual	~GlobalLinSys ()

const GlobalLinSysKey &	GetKey (void) const
	Returns the key associated with the system. More...

const boost::weak_ptr< ExpList > &	GetLocMat (void) const

void	InitObject ()

void	Initialise (const boost::shared_ptr< AssemblyMap > &pLocToGloMap)

void	Solve (const Array< OneD, const NekDouble > &in, Array< OneD, NekDouble > &out, const AssemblyMapSharedPtr &locToGloMap, const Array< OneD, const NekDouble > &dirForcing=NullNekDouble1DArray)
	Solve the linear system for given input and output vectors using a specified local to global map. More...

boost::shared_ptr< GlobalLinSys >	GetSharedThisPtr ()
	Returns a shared pointer to the current object. More...

int	GetNumBlocks ()

DNekScalMatSharedPtr	GetBlock (unsigned int n)

DNekScalBlkMatSharedPtr	GetStaticCondBlock (unsigned int n)

void	DropStaticCondBlock (unsigned int n)

void	SolveLinearSystem (const int pNumRows, const Array< OneD, const NekDouble > &pInput, Array< OneD, NekDouble > &pOutput, const AssemblyMapSharedPtr &locToGloMap, const int pNumDir=0)
	Solve the linear system for given input and output vectors. More...

Protected Member Functions
virtual DNekScalBlkMatSharedPtr	v_PreSolve (int scLevel, NekVector< NekDouble > &F_GlobBnd)

virtual void	v_BasisTransform (Array< OneD, NekDouble > &pInOut, int offset)

virtual void	v_BasisInvTransform (Array< OneD, NekDouble > &pInOut)

virtual void	v_AssembleSchurComplement (boost::shared_ptr< AssemblyMap > pLoctoGloMap)

virtual int	v_GetNumBlocks ()
	Get the number of blocks in this system. More...

virtual GlobalLinSysStaticCondSharedPtr	v_Recurse (const GlobalLinSysKey &mkey, const boost::weak_ptr< ExpList > &pExpList, const DNekScalBlkMatSharedPtr pSchurCompl, const DNekScalBlkMatSharedPtr pBinvD, const DNekScalBlkMatSharedPtr pC, const DNekScalBlkMatSharedPtr pInvD, const boost::shared_ptr< AssemblyMap > &locToGloMap)=0

virtual void	v_Solve (const Array< OneD, const NekDouble > &in, Array< OneD, NekDouble > &out, const AssemblyMapSharedPtr &locToGloMap, const Array< OneD, const NekDouble > &dirForcing=NullNekDouble1DArray)
	Solve the linear system for given input and output vectors using a specified local to global map. More...

virtual void	v_InitObject ()

virtual void	v_Initialise (const boost::shared_ptr< AssemblyMap > &locToGloMap)
	Initialise this object. More...

void	SetupTopLevel (const boost::shared_ptr< AssemblyMap > &locToGloMap)
	Set up the storage for the Schur complement or the top level of the multi-level Schur complement. More...

void	ConstructNextLevelCondensedSystem (const boost::shared_ptr< AssemblyMap > &locToGloMap)

Protected Member Functions inherited from Nektar::MultiRegions::GlobalLinSys
virtual DNekScalMatSharedPtr	v_GetBlock (unsigned int n)
	Retrieves the block matrix from n-th expansion using the matrix key provided by the m_linSysKey. More...

virtual DNekScalBlkMatSharedPtr	v_GetStaticCondBlock (unsigned int n)
	Retrieves a the static condensation block matrices from n-th expansion using the matrix key provided by the m_linSysKey. More...

virtual void	v_DropStaticCondBlock (unsigned int n)
	Releases the static condensation block matrices from NekManager of n-th expansion using the matrix key provided by the m_linSysKey. More...

PreconditionerSharedPtr	CreatePrecon (AssemblyMapSharedPtr asmMap)
	Create a preconditioner object from the parameters defined in the supplied assembly map. More...

Protected Attributes
GlobalLinSysStaticCondSharedPtr	m_recursiveSchurCompl
	Schur complement for Direct Static Condensation. More...

DNekScalBlkMatSharedPtr	m_schurCompl
	Block Schur complement matrix. More...

DNekScalBlkMatSharedPtr	m_BinvD
	Block $BD^{-1}$ matrix. More...

DNekScalBlkMatSharedPtr	m_C
	Block matrix. More...

DNekScalBlkMatSharedPtr	m_invD
	Block $D^{-1}$ matrix. More...

boost::shared_ptr< AssemblyMap >	m_locToGloMap
	Local to global map. More...

Array< OneD, NekDouble >	m_wsp
	Workspace array for matrix multiplication. More...

Protected Attributes inherited from Nektar::MultiRegions::GlobalLinSys
const GlobalLinSysKey	m_linSysKey
	Key associated with this linear system. More...

const boost::weak_ptr< ExpList >	m_expList
	Local Matrix System. More...

const map< int, RobinBCInfoSharedPtr >	m_robinBCInfo
	Robin boundary info. More...

Detailed Description

A global linear system.

Solves a linear system using single- or multi-level static condensation.

Definition at line 55 of file GlobalLinSysStaticCond.h.

Constructor & Destructor Documentation

Nektar::MultiRegions::GlobalLinSysStaticCond::GlobalLinSysStaticCond	(	const GlobalLinSysKey &	pKey,
		const boost::weak_ptr< ExpList > &	pExpList,
		const boost::shared_ptr< AssemblyMap > &	pLocToGloMap
	)

Constructor for full direct matrix solve.

For a matrix system of the form

$\left[ \begin{array}{cc} \boldsymbol{A} & \boldsymbol{B}\\ \boldsymbol{C} & \boldsymbol{D} \end{array} \right] \left[ \begin{array}{c} \boldsymbol{x_1}\\ \boldsymbol{x_2} \end{array}\right] = \left[ \begin{array}{c} \boldsymbol{y_1}\\ \boldsymbol{y_2} \end{array}\right],$

where $\boldsymbol{D}$ and $(\boldsymbol{A-BD^{-1}C})$ are invertible, store and assemble a static condensation system, according to a given local to global mapping. #m_linSys is constructed by AssembleSchurComplement().

Parameters

mKey	Associated matrix key.
pLocMatSys	LocalMatrixSystem
locToGloMap	Local to global mapping.

Definition at line 74 of file GlobalLinSysStaticCond.cpp.

                 : GlobalLinSys(pKey, pExpList, pLocToGloMap),
                   m_locToGloMap (pLocToGloMap)
         {
         }

Nektar::MultiRegions::GlobalLinSysStaticCond::~GlobalLinSysStaticCond ( )

virtual

Definition at line 95 of file GlobalLinSysStaticCond.cpp.

         {
 
         }

Member Function Documentation

void Nektar::MultiRegions::GlobalLinSysStaticCond::ConstructNextLevelCondensedSystem ( const boost::shared_ptr< AssemblyMap > & locToGloMap )

protected

Definition at line 344 of file GlobalLinSysStaticCond.cpp.

References Nektar::MemoryManager< DataType >::AllocateSharedPtr(), ASSERTL0, Nektar::eDIAGONAL, Nektar::eWrapper, Nektar::MultiRegions::GlobalLinSys::m_expList, Nektar::MultiRegions::GlobalLinSys::m_linSysKey, m_recursiveSchurCompl, m_schurCompl, sign, and v_Recurse().

Referenced by v_Initialise().

         {
             int i,j,n,cnt;
             DNekScalBlkMatSharedPtr blkMatrices[4];
 
             // Create temporary matrices within an inner-local scope to ensure
             // any references to the intermediate storage is lost before
             // the recursive step, rather than at the end of the routine.
             // This allows the schur complement matrix from this level to be
             // disposed of in the next level after use without being retained
             // due to lingering shared pointers.
             {
 
                 const Array<OneD,const unsigned int>& nBndDofsPerPatch
                                 = pLocToGloMap->GetNumLocalBndCoeffsPerPatch();
                 const Array<OneD,const unsigned int>& nIntDofsPerPatch
                                 = pLocToGloMap->GetNumLocalIntCoeffsPerPatch();
 
                 // STEP 1:
                 // Based upon the schur complement of the the current level we
                 // will substructure this matrix in the form
                 //      --     --
                 //      | A   B |
                 //      | C   D |
                 //      --     --
                 // All matrices A,B,C and D are (diagonal) blockmatrices.
                 // However, as a start we will use an array of DNekMatrices as
                 // it is too hard to change the individual entries of a
                 // DNekScalBlkMatSharedPtr.
 
                 // In addition, we will also try to ensure that the memory of
                 // the blockmatrices will be contiguous. This will probably
                 // enhance the efficiency
                 // - Calculate the total number of entries in the blockmatrices
                 int nPatches  = pLocToGloMap->GetNumPatches();
                 int nEntriesA = 0; int nEntriesB = 0;
                 int nEntriesC = 0; int nEntriesD = 0;
 
                 for(i = 0; i < nPatches; i++)
                 {
                     nEntriesA += nBndDofsPerPatch[i]*nBndDofsPerPatch[i];
                     nEntriesB += nBndDofsPerPatch[i]*nIntDofsPerPatch[i];
                     nEntriesC += nIntDofsPerPatch[i]*nBndDofsPerPatch[i];
                     nEntriesD += nIntDofsPerPatch[i]*nIntDofsPerPatch[i];
                 }
 
                 // Now create the DNekMatrices and link them to the memory
                 // allocated above
                 Array<OneD, DNekMatSharedPtr> substructuredMat[4]
                     = {Array<OneD, DNekMatSharedPtr>(nPatches),  //Matrix A
                        Array<OneD, DNekMatSharedPtr>(nPatches),  //Matrix B
                        Array<OneD, DNekMatSharedPtr>(nPatches),  //Matrix C
                        Array<OneD, DNekMatSharedPtr>(nPatches)}; //Matrix D
 
                 // Initialise storage for the matrices. We do this separately
                 // for each matrix so the matrices may be independently
                 // deallocated when no longer required.
                 Array<OneD, NekDouble> storageA(nEntriesA,0.0);
                 Array<OneD, NekDouble> storageB(nEntriesB,0.0);
                 Array<OneD, NekDouble> storageC(nEntriesC,0.0);
                 Array<OneD, NekDouble> storageD(nEntriesD,0.0);
 
                 Array<OneD, NekDouble> tmparray;
                 PointerWrapper wType = eWrapper;
                 int cntA = 0;
                 int cntB = 0;
                 int cntC = 0;
                 int cntD = 0;
 
                 for(i = 0; i < nPatches; i++)
                 {
                     // Matrix A
                     tmparray = storageA+cntA;
                     substructuredMat[0][i] = MemoryManager<DNekMat>
                                     ::AllocateSharedPtr(nBndDofsPerPatch[i],
                                                         nBndDofsPerPatch[i],
                                                         tmparray, wType);
                     // Matrix B
                     tmparray = storageB+cntB;
                     substructuredMat[1][i] = MemoryManager<DNekMat>
                                     ::AllocateSharedPtr(nBndDofsPerPatch[i],
                                                         nIntDofsPerPatch[i],
                                                         tmparray, wType);
                     // Matrix C
                     tmparray = storageC+cntC;
                     substructuredMat[2][i] = MemoryManager<DNekMat>
                                     ::AllocateSharedPtr(nIntDofsPerPatch[i],
                                                         nBndDofsPerPatch[i],
                                                         tmparray, wType);
                     // Matrix D
                     tmparray = storageD+cntD;
                     substructuredMat[3][i] = MemoryManager<DNekMat>
                                     ::AllocateSharedPtr(nIntDofsPerPatch[i],
                                                         nIntDofsPerPatch[i],
                                                         tmparray, wType);
 
                     cntA += nBndDofsPerPatch[i] * nBndDofsPerPatch[i];
                     cntB += nBndDofsPerPatch[i] * nIntDofsPerPatch[i];
                     cntC += nIntDofsPerPatch[i] * nBndDofsPerPatch[i];
                     cntD += nIntDofsPerPatch[i] * nIntDofsPerPatch[i];
                 }
 
                 // Then, project SchurComplement onto
                 // the substructured matrices of the next level
                 DNekScalBlkMatSharedPtr SchurCompl  = m_schurCompl;
                 DNekScalMatSharedPtr schurComplSubMat;
                 int       schurComplSubMatnRows;
                 Array<OneD, const int>       patchId, dofId;
                 Array<OneD, const unsigned int>      isBndDof;
                 Array<OneD, const NekDouble> sign;
                 NekDouble scale;
 
                 for(n = cnt = 0; n < SchurCompl->GetNumberOfBlockRows(); ++n)
                 {
                     schurComplSubMat      = SchurCompl->GetBlock(n,n);
                     schurComplSubMatnRows = schurComplSubMat->GetRows();
                     
                     scale = SchurCompl->GetBlock(n,n)->Scale();
                     Array<OneD, NekDouble> schurSubMat
                         = SchurCompl->GetBlock(n,n)->GetOwnedMatrix()->GetPtr();
                     
                     patchId  = pLocToGloMap->GetPatchMapFromPrevLevel()
                                ->GetPatchId() + cnt;
                     dofId    = pLocToGloMap->GetPatchMapFromPrevLevel()
                                ->GetDofId()   + cnt;
                     isBndDof = pLocToGloMap->GetPatchMapFromPrevLevel()
                                ->IsBndDof() + cnt;
                     sign     = pLocToGloMap->GetPatchMapFromPrevLevel()
                                ->GetSign() + cnt;
 
                     // Set up  Matrix;
                     for(i = 0; i < schurComplSubMatnRows; ++i)
                     {
                         int pId = patchId[i];
                         Array<OneD, NekDouble> subMat0
                             = substructuredMat[0][pId]->GetPtr();
                         Array<OneD, NekDouble> subMat1
                             = substructuredMat[1][patchId[i]]->GetPtr();
                         Array<OneD, NekDouble> subMat2
                             = substructuredMat[2][patchId[i]]->GetPtr();
                         Array<OneD, NekDouble> subMat3
                             = substructuredMat[3][patchId[i]]->GetPtr();
                         int subMat0rows = substructuredMat[0][pId]->GetRows();
                         int subMat1rows = substructuredMat[1][pId]->GetRows();
                         int subMat2rows = substructuredMat[2][pId]->GetRows();
                         int subMat3rows = substructuredMat[3][pId]->GetRows();
                         
                         if(isBndDof[i])
                         {
                             for(j = 0; j < schurComplSubMatnRows; ++j)
                             {
                                 ASSERTL0(patchId[i]==patchId[j],
                                          "These values should be equal");
                                 
                                 if(isBndDof[j])
                                 {
                                     subMat0[dofId[i]+dofId[j]*subMat0rows] +=
                                         sign[i]*sign[j]*(
                                             scale*schurSubMat[
                                                 i+j*schurComplSubMatnRows]);
                                 }
                                 else
                                 {
                                     subMat1[dofId[i]+dofId[j]*subMat1rows] +=
                                         sign[i]*sign[j]*(
                                             scale*schurSubMat[
                                                 i+j*schurComplSubMatnRows]);
                                 }
                             }
                         }
                         else
                         {
                             for(j = 0; j < schurComplSubMatnRows; ++j)
                             {
                                 ASSERTL0(patchId[i]==patchId[j],
                                          "These values should be equal");
                                 
                                 if(isBndDof[j])
                                 {
                                     subMat2[dofId[i]+dofId[j]*subMat2rows] +=
                                         sign[i]*sign[j]*(
                                             scale*schurSubMat[
                                                 i+j*schurComplSubMatnRows]);
                                 }
                                 else
                                 {
                                     subMat3[dofId[i]+dofId[j]*subMat3rows] +=
                                         sign[i]*sign[j]*(
                                             scale*schurSubMat[
                                                 i+j*schurComplSubMatnRows]);
                                 }
                             }
                         }
                     }
                     cnt += schurComplSubMatnRows;
                 }
 
                 // STEP 2: condense the system
                 // This can be done elementally (i.e. patch per patch)
                 for(i = 0; i < nPatches; i++)
                 {
                     if(nIntDofsPerPatch[i])
                     {
                         Array<OneD, NekDouble> subMat0
                             = substructuredMat[0][i]->GetPtr();
                         Array<OneD, NekDouble> subMat1
                             = substructuredMat[1][i]->GetPtr();
                         Array<OneD, NekDouble> subMat2
                             = substructuredMat[2][i]->GetPtr();
                         int subMat0rows = substructuredMat[0][i]->GetRows();
                         int subMat1rows = substructuredMat[1][i]->GetRows();
                         int subMat2rows = substructuredMat[2][i]->GetRows();
                         int subMat2cols = substructuredMat[2][i]->GetColumns();
 
                         // 1. D -> InvD
                         substructuredMat[3][i]->Invert();
                         // 2. B -> BInvD
                         (*substructuredMat[1][i]) = (*substructuredMat[1][i])*
                             (*substructuredMat[3][i]);
                         // 3. A -> A - BInvD*C (= schurcomplement)
                         // (*substructuredMat[0][i]) =(*substructuredMat[0][i])-
                         // (*substructuredMat[1][i])*(*substructuredMat[2][i]);
                         // Note: faster to use blas directly
                         Blas::Dgemm('N','N', subMat1rows, subMat2cols,
                                     subMat2rows, -1.0, &subMat1[0], subMat1rows,
                                     &subMat2[0], subMat2rows, 1.0, &subMat0[0],
                                     subMat0rows);
                     }
                 }
 
                 // STEP 3: fill the block matrices. However, do note that we
                 // first have to convert them to a DNekScalMat in order to be
                 // compatible with the first level of static condensation
                 const Array<OneD,const unsigned int>& nbdry_size
                     = pLocToGloMap->GetNumLocalBndCoeffsPerPatch();
                 const Array<OneD,const unsigned int>& nint_size
                     = pLocToGloMap->GetNumLocalIntCoeffsPerPatch();
                 MatrixStorage blkmatStorage = eDIAGONAL;
 
                 blkMatrices[0] = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nbdry_size, nbdry_size, blkmatStorage);
                 blkMatrices[1] = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nbdry_size, nint_size , blkmatStorage);
                 blkMatrices[2] = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nint_size , nbdry_size, blkmatStorage);
                 blkMatrices[3] = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nint_size , nint_size , blkmatStorage);
 
                 DNekScalMatSharedPtr tmpscalmat;
                 for(i = 0; i < nPatches; i++)
                 {
                     for(j = 0; j < 4; j++)
                     {
                         tmpscalmat= MemoryManager<DNekScalMat>
                             ::AllocateSharedPtr(1.0,substructuredMat[j][i]);
                         blkMatrices[j]->SetBlock(i,i,tmpscalmat);
                     }
                 }
             }
 
             // We've finished with the Schur complement matrix passed to this
             // level, so return the memory to the system. The Schur complement
             // matrix need only be retained at the last level. Save the other
             // matrices at this level though.
             m_schurCompl.reset();
 
             m_recursiveSchurCompl = v_Recurse(
                 m_linSysKey, m_expList, blkMatrices[0], blkMatrices[1],
                 blkMatrices[2], blkMatrices[3], pLocToGloMap);
         }

void Nektar::MultiRegions::GlobalLinSysStaticCond::SetupTopLevel ( const boost::shared_ptr< AssemblyMap > & pLocToGloMap )

protected

Set up the storage for the Schur complement or the top level of the multi-level Schur complement.

For the first level in multi-level static condensation, or the only level in the case of single-level static condensation, allocate the condensed matrices and populate them with the local matrices retrieved from the expansion list.

Parameters

Definition at line 295 of file GlobalLinSysStaticCond.cpp.

References Nektar::MemoryManager< DataType >::AllocateSharedPtr(), Nektar::eDIAGONAL, Nektar::StdRegions::eHybridDGHelmBndLam, Nektar::MultiRegions::GlobalMatrixKey::GetMatrixType(), m_BinvD, m_C, Nektar::MultiRegions::GlobalLinSys::m_expList, m_invD, Nektar::MultiRegions::GlobalLinSys::m_linSysKey, m_schurCompl, Nektar::MultiRegions::GlobalLinSys::v_GetBlock(), and Nektar::MultiRegions::GlobalLinSys::v_GetStaticCondBlock().

Referenced by Nektar::MultiRegions::GlobalLinSysPETScStaticCond::v_InitObject(), v_InitObject(), and Nektar::MultiRegions::GlobalLinSysIterativeStaticCond::v_InitObject().

         {
             int n;
             int n_exp = m_expList.lock()->GetNumElmts();
 
             const Array<OneD,const unsigned int>& nbdry_size
                     = pLocToGloMap->GetNumLocalBndCoeffsPerPatch();
             const Array<OneD,const unsigned int>& nint_size
                     = pLocToGloMap->GetNumLocalIntCoeffsPerPatch();
 
             // Setup Block Matrix systems
             MatrixStorage blkmatStorage = eDIAGONAL;
             m_schurCompl = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nbdry_size, nbdry_size, blkmatStorage);
             m_BinvD      = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nbdry_size, nint_size , blkmatStorage);
             m_C          = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nint_size , nbdry_size, blkmatStorage);
             m_invD       = MemoryManager<DNekScalBlkMat>
                     ::AllocateSharedPtr(nint_size , nint_size , blkmatStorage);
 
             for(n = 0; n < n_exp; ++n)
             {
                 if (m_linSysKey.GetMatrixType() ==
                         StdRegions::eHybridDGHelmBndLam)
                 {
                     DNekScalMatSharedPtr loc_mat
                         = GlobalLinSys::v_GetBlock(
                             m_expList.lock()->GetOffset_Elmt_Id(n));
                     m_schurCompl->SetBlock(n,n,loc_mat);
                 }
                 else
                 {
                     DNekScalBlkMatSharedPtr loc_schur
                         = GlobalLinSys::v_GetStaticCondBlock(
                             m_expList.lock()->GetOffset_Elmt_Id(n));
                     DNekScalMatSharedPtr t;
                     m_schurCompl->SetBlock(n, n, t = loc_schur->GetBlock(0,0));
                     m_BinvD     ->SetBlock(n, n, t = loc_schur->GetBlock(0,1));
                     m_C         ->SetBlock(n, n, t = loc_schur->GetBlock(1,0));
                     m_invD      ->SetBlock(n, n, t = loc_schur->GetBlock(1,1));
                 }
             }
         }

virtual void Nektar::MultiRegions::GlobalLinSysStaticCond::v_AssembleSchurComplement ( boost::shared_ptr< AssemblyMap > pLoctoGloMap )

inlineprotectedvirtual

Reimplemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, Nektar::MultiRegions::GlobalLinSysXxtStaticCond, Nektar::MultiRegions::GlobalLinSysPETScStaticCond, and Nektar::MultiRegions::GlobalLinSysDirectStaticCond.

Definition at line 87 of file GlobalLinSysStaticCond.h.

Referenced by v_Initialise().

             {
                 
             }

virtual void Nektar::MultiRegions::GlobalLinSysStaticCond::v_BasisInvTransform ( Array< OneD, NekDouble > & pInOut )

inlineprotectedvirtual

Reimplemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, and Nektar::MultiRegions::GlobalLinSysPETScStaticCond.

Definition at line 81 of file GlobalLinSysStaticCond.h.

Referenced by v_Solve().

             {
                 
             }

virtual void Nektar::MultiRegions::GlobalLinSysStaticCond::v_BasisTransform	(	Array< OneD, NekDouble > &	pInOut,
		int	offset
	)

inlineprotectedvirtual

Reimplemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, and Nektar::MultiRegions::GlobalLinSysPETScStaticCond.

Definition at line 74 of file GlobalLinSysStaticCond.h.

Referenced by v_Solve().

             {
 
             }

int Nektar::MultiRegions::GlobalLinSysStaticCond::v_GetNumBlocks ( )

protectedvirtual

Get the number of blocks in this system.

At the top level this corresponds to the number of elements in the expansion list.

Reimplemented from Nektar::MultiRegions::GlobalLinSys.

Definition at line 283 of file GlobalLinSysStaticCond.cpp.

References m_schurCompl.

         {
             return m_schurCompl->GetNumberOfBlockRows();
         }

void Nektar::MultiRegions::GlobalLinSysStaticCond::v_Initialise ( const boost::shared_ptr< AssemblyMap > & pLocToGloMap )

protectedvirtual

Initialise this object.

If at the last level of recursion (or the only level in the case of single-level static condensation), assemble the Schur complement. For other levels, in the case of multi-level static condensation, the next level of the condensed system is computed.

Parameters

pLocToGloMap Local to global mapping.

Reimplemented from Nektar::MultiRegions::GlobalLinSys.

Definition at line 265 of file GlobalLinSysStaticCond.cpp.

References ConstructNextLevelCondensedSystem(), m_locToGloMap, m_wsp, and v_AssembleSchurComplement().

         {
             int nLocalBnd = m_locToGloMap->GetNumLocalBndCoeffs();
             int nGlobal = m_locToGloMap->GetNumGlobalCoeffs();
             m_wsp = Array<OneD, NekDouble>(2*nLocalBnd + nGlobal, 0.0);
 
             if (pLocToGloMap->AtLastLevel())
             {
                 v_AssembleSchurComplement(pLocToGloMap);
             }
             else
             {
                 ConstructNextLevelCondensedSystem(
                         pLocToGloMap->GetNextLevelLocalToGlobalMap());
             }
         }

void Nektar::MultiRegions::GlobalLinSysStaticCond::v_InitObject ( )

protectedvirtual

Reimplemented from Nektar::MultiRegions::GlobalLinSys.

Reimplemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, and Nektar::MultiRegions::GlobalLinSysPETScStaticCond.

Definition at line 83 of file GlobalLinSysStaticCond.cpp.

References Nektar::MultiRegions::GlobalLinSys::Initialise(), m_locToGloMap, and SetupTopLevel().

         {
             // Allocate memory for top-level structure
             SetupTopLevel(m_locToGloMap);
 
             // Construct this level
             Initialise(m_locToGloMap);
         }

virtual DNekScalBlkMatSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::v_PreSolve	(	int	scLevel,
		NekVector< NekDouble > &	F_GlobBnd
	)

inlineprotectedvirtual

Reimplemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, and Nektar::MultiRegions::GlobalLinSysPETScStaticCond.

Definition at line 67 of file GlobalLinSysStaticCond.h.

References m_schurCompl.

Referenced by v_Solve().

             {
                 return m_schurCompl;
             }

virtual GlobalLinSysStaticCondSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::v_Recurse	(	const GlobalLinSysKey &	mkey,
		const boost::weak_ptr< ExpList > &	pExpList,
		const DNekScalBlkMatSharedPtr	pSchurCompl,
		const DNekScalBlkMatSharedPtr	pBinvD,
		const DNekScalBlkMatSharedPtr	pC,
		const DNekScalBlkMatSharedPtr	pInvD,
		const boost::shared_ptr< AssemblyMap > &	locToGloMap
	)

protectedpure virtual

Implemented in Nektar::MultiRegions::GlobalLinSysIterativeStaticCond, Nektar::MultiRegions::GlobalLinSysPETScStaticCond, Nektar::MultiRegions::GlobalLinSysXxtStaticCond, and Nektar::MultiRegions::GlobalLinSysDirectStaticCond.

Referenced by ConstructNextLevelCondensedSystem().

void Nektar::MultiRegions::GlobalLinSysStaticCond::v_Solve	(	const Array< OneD, const NekDouble > &	in,
		Array< OneD, NekDouble > &	out,
		const AssemblyMapSharedPtr &	locToGloMap,
		const Array< OneD, const NekDouble > &	dirForcing = `NullNekDouble1DArray`
	)

protectedvirtual

Solve the linear system for given input and output vectors using a specified local to global map.

Implements Nektar::MultiRegions::GlobalLinSys.

Definition at line 104 of file GlobalLinSysStaticCond.cpp.

References Nektar::eWrapper, Vmath::Fill(), Nektar::NekVector< DataType >::GetPtr(), m_BinvD, m_C, m_invD, m_recursiveSchurCompl, m_wsp, Nektar::MultiRegions::GlobalLinSys::SolveLinearSystem(), v_BasisInvTransform(), v_BasisTransform(), v_PreSolve(), Vmath::Vadd(), Vmath::Vcopy(), and Vmath::Vsub().

         {
             bool dirForcCalculated = (bool) dirForcing.num_elements();
             bool atLastLevel       = pLocToGloMap->AtLastLevel();
             int  scLevel           = pLocToGloMap->GetStaticCondLevel();
             
             int nGlobDofs          = pLocToGloMap->GetNumGlobalCoeffs();
             int nGlobBndDofs       = pLocToGloMap->GetNumGlobalBndCoeffs();
             int nDirBndDofs        = pLocToGloMap->GetNumGlobalDirBndCoeffs();
             int nGlobHomBndDofs    = nGlobBndDofs - nDirBndDofs;
             int nLocBndDofs        = pLocToGloMap->GetNumLocalBndCoeffs();
             int nIntDofs           = pLocToGloMap->GetNumGlobalCoeffs()
                 - nGlobBndDofs;
 
             Array<OneD, NekDouble> F = m_wsp + 2*nLocBndDofs;
             Array<OneD, NekDouble> tmp;
             if(nDirBndDofs && dirForcCalculated)
             {
                 Vmath::Vsub(nGlobDofs,in.get(),1,dirForcing.get(),1,F.get(),1);
             }
             else
             {
                 Vmath::Vcopy(nGlobDofs,in.get(),1,F.get(),1);
             }
             
             NekVector<NekDouble> F_HomBnd(nGlobHomBndDofs,tmp=F+nDirBndDofs,
                                           eWrapper);
             NekVector<NekDouble> F_GlobBnd(nGlobBndDofs,F,eWrapper);
             NekVector<NekDouble> F_LocBnd(nLocBndDofs,0.0);
             NekVector<NekDouble> F_Int(nIntDofs,tmp=F+nGlobBndDofs,eWrapper);
             
             NekVector<NekDouble> V_GlobBnd(nGlobBndDofs,out,eWrapper);
             NekVector<NekDouble> V_GlobHomBnd(nGlobHomBndDofs,
                                               tmp=out+nDirBndDofs,
                                               eWrapper);
             NekVector<NekDouble> V_Int(nIntDofs,tmp=out+nGlobBndDofs,eWrapper);
             NekVector<NekDouble> V_LocBnd(nLocBndDofs,m_wsp,eWrapper);
             
             NekVector<NekDouble> V_GlobHomBndTmp(nGlobHomBndDofs,0.0);
 
             // set up normalisation factor for right hand side on first SC level
             DNekScalBlkMatSharedPtr sc = v_PreSolve(scLevel, F_GlobBnd);
 
             if(nGlobHomBndDofs)
             {
                 // construct boundary forcing
                 if( nIntDofs  && ((!dirForcCalculated) && (atLastLevel)) )
                 {
                     DNekScalBlkMat &BinvD      = *m_BinvD;
                     DNekScalBlkMat &SchurCompl = *sc;
 
                     // include dirichlet boundary forcing
                     pLocToGloMap->GlobalToLocalBnd(V_GlobBnd,V_LocBnd);
                     V_LocBnd = BinvD*F_Int + SchurCompl*V_LocBnd;
                 }
                 else if((!dirForcCalculated) && (atLastLevel))
                 {
                     // include dirichlet boundary forcing
                     DNekScalBlkMat &SchurCompl = *sc;
                     pLocToGloMap->GlobalToLocalBnd(V_GlobBnd,V_LocBnd);
                     V_LocBnd = SchurCompl*V_LocBnd;
                 }
                 else
                 {
                     DNekScalBlkMat &BinvD      = *m_BinvD;
                     V_LocBnd = BinvD*F_Int;
                 }
                 
                 pLocToGloMap->AssembleBnd(V_LocBnd,V_GlobHomBndTmp,
                                           nDirBndDofs);
                 F_HomBnd = F_HomBnd - V_GlobHomBndTmp;
 
                 // Transform from original basis to low energy
                 v_BasisTransform(F, nDirBndDofs);
 
                 // For parallel multi-level static condensation some
                 // processors may have different levels to others. This
                 // routine receives contributions to partition vertices from
                 // those lower levels, whilst not sending anything to the
                 // other partitions, and includes them in the modified right
                 // hand side vector.
                 int lcLevel = pLocToGloMap->GetLowestStaticCondLevel();
                 if(atLastLevel && scLevel < lcLevel)
                 {
                     // If this level is not the lowest level across all
                     // processes, we must do dummy communication for the
                     // remaining levels
                     Array<OneD, NekDouble> tmp(nGlobBndDofs);
                     for (int i = scLevel; i < lcLevel; ++i)
                     {
                         Vmath::Fill(nGlobBndDofs, 0.0, tmp, 1);
                         pLocToGloMap->UniversalAssembleBnd(tmp);
                         Vmath::Vcopy(nGlobHomBndDofs,
                                      tmp.get()+nDirBndDofs,          1,
                                      V_GlobHomBndTmp.GetPtr().get(), 1);
                         F_HomBnd = F_HomBnd - V_GlobHomBndTmp;
                     }
                 }
 
                 // solve boundary system
                 if(atLastLevel)
                 {
                     Array<OneD, NekDouble> pert(nGlobBndDofs,0.0);
                     NekVector<NekDouble>   Pert(nGlobBndDofs,pert,eWrapper);
 
                     // Solve for difference from initial solution given inout;
                     SolveLinearSystem(
                         nGlobBndDofs, F, pert, pLocToGloMap, nDirBndDofs);
 
                     // Transform back to original basis
                     v_BasisInvTransform(pert);
 
                     // Add back initial conditions onto difference
                     Vmath::Vadd(nGlobHomBndDofs,&out[nDirBndDofs],1,
                                 &pert[nDirBndDofs],1,&out[nDirBndDofs],1);
                 }
                 else
                 {
                     m_recursiveSchurCompl->Solve(F,
                                 V_GlobBnd.GetPtr(),
                                 pLocToGloMap->GetNextLevelLocalToGlobalMap());
                 }
             }
 
             // solve interior system
             if(nIntDofs)
             {
                 DNekScalBlkMat &invD  = *m_invD;
 
                 if(nGlobHomBndDofs || nDirBndDofs)
                 {
                     DNekScalBlkMat &C     = *m_C;
 
                     if(dirForcCalculated && nDirBndDofs)
                     {
                         pLocToGloMap->GlobalToLocalBnd(V_GlobHomBnd,V_LocBnd,
                                                       nDirBndDofs);
                     }
                     else
                     {
                         pLocToGloMap->GlobalToLocalBnd(V_GlobBnd,V_LocBnd);
                     }
                     F_Int = F_Int - C*V_LocBnd;
                 }
 
                 V_Int = invD*F_Int;
             }
         }

Member Data Documentation

DNekScalBlkMatSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::m_BinvD

protected

Block $BD^{-1}$ matrix.

Definition at line 109 of file GlobalLinSysStaticCond.h.

Referenced by Nektar::MultiRegions::GlobalLinSysDirectStaticCond::GlobalLinSysDirectStaticCond(), Nektar::MultiRegions::GlobalLinSysIterativeStaticCond::GlobalLinSysIterativeStaticCond(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::GlobalLinSysPETScStaticCond(), Nektar::MultiRegions::GlobalLinSysXxtStaticCond::GlobalLinSysXxtStaticCond(), SetupTopLevel(), Nektar::MultiRegions::GlobalLinSysDirectStaticCond::v_AssembleSchurComplement(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::v_AssembleSchurComplement(), and v_Solve().

DNekScalBlkMatSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::m_C

protected

Block $ C $ matrix.

Definition at line 111 of file GlobalLinSysStaticCond.h.

Referenced by Nektar::MultiRegions::GlobalLinSysDirectStaticCond::GlobalLinSysDirectStaticCond(), Nektar::MultiRegions::GlobalLinSysIterativeStaticCond::GlobalLinSysIterativeStaticCond(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::GlobalLinSysPETScStaticCond(), Nektar::MultiRegions::GlobalLinSysXxtStaticCond::GlobalLinSysXxtStaticCond(), SetupTopLevel(), Nektar::MultiRegions::GlobalLinSysDirectStaticCond::v_AssembleSchurComplement(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::v_AssembleSchurComplement(), and v_Solve().

DNekScalBlkMatSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::m_invD

protected

Block $D^{-1}$ matrix.

Definition at line 113 of file GlobalLinSysStaticCond.h.

Referenced by Nektar::MultiRegions::GlobalLinSysDirectStaticCond::GlobalLinSysDirectStaticCond(), Nektar::MultiRegions::GlobalLinSysIterativeStaticCond::GlobalLinSysIterativeStaticCond(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::GlobalLinSysPETScStaticCond(), Nektar::MultiRegions::GlobalLinSysXxtStaticCond::GlobalLinSysXxtStaticCond(), SetupTopLevel(), Nektar::MultiRegions::GlobalLinSysDirectStaticCond::v_AssembleSchurComplement(), Nektar::MultiRegions::GlobalLinSysPETScStaticCond::v_AssembleSchurComplement(), and v_Solve().

boost::shared_ptr<AssemblyMap> Nektar::MultiRegions::GlobalLinSysStaticCond::m_locToGloMap

protected

Local to global map.

Definition at line 115 of file GlobalLinSysStaticCond.h.

GlobalLinSysStaticCondSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::m_recursiveSchurCompl

protected

Schur complement for Direct Static Condensation.

Definition at line 105 of file GlobalLinSysStaticCond.h.

Referenced by ConstructNextLevelCondensedSystem(), and v_Solve().

DNekScalBlkMatSharedPtr Nektar::MultiRegions::GlobalLinSysStaticCond::m_schurCompl

protected

Block Schur complement matrix.

Definition at line 107 of file GlobalLinSysStaticCond.h.

Array<OneD, NekDouble> Nektar::MultiRegions::GlobalLinSysStaticCond::m_wsp

protected

Workspace array for matrix multiplication.

Definition at line 117 of file GlobalLinSysStaticCond.h.

Referenced by Nektar::MultiRegions::GlobalLinSysIterativeStaticCond::v_DoMatrixMultiply(), v_Initialise(), and v_Solve().

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation