PARDISO Advanced Initialization VI

Owning Palette: Sparse Linear Algebra VIs

Requires: Multicore Analysis and Sparse Matrix Toolkit

Initializes a PARDISO session, attaches a sparse matrix to the initialized session, and analyzes and performs symbolic factorization on the sparse matrix. The PARDISO Advanced Factorization VI uses the analysis result for numerical factorization.

Details  

PARDISO Advanced Initialization (DBL)

	Input Matrix specifies a square sparse matrix to analyze and perform symbolic factorization.
	matrix type specifies the matrix type of the Input Matrix. 0 General (default)—Specifies a square matrix. 1 Symmetric—Specifies a symmetric matrix. LabVIEW only uses the upper triangular part in the Input Matrix to solve the linear system. 2 Hermitian—Specifies a Hermitian matrix. This option is the same as Symmetric when the Input Matrix is a real-valued matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero. 3 Positive Definite—Specifies a symmetric positive definite or Hermitian positive definite matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero.
	error in describes error conditions that occur before this node runs. This input provides standard error in functionality.
	Solver Parameters specifies the set of parameters to use when configuring PARDISO. ordering method specifies the method of permutation in order to reduce fill-in on factorization. 0 Minimum Degree—Specifies to use the minimum degree algorithm. 1 Nested Dissection (default)—Specifies to use the nested dissection algorithm. 2 Parallel Nested Dissection—Specifies to use the parallel version of nested dissection algorithm. 3 User Defined Permutation—Specifies to use the User Defined Permutation as the permutation vector to reduce fill-in. User Defined Permutation specifies the permutation vector to reduce fill-in. LabVIEW uses the user defined permutation vector only when ordering method is User Defined Permutation. output permutation? specifies whether LabVIEW returns the permutation vector that PARDISO actually used. The default is FALSE. max # of iterative refinement specifies the maximum number of iterative refinement that PARDISO can perform. The default is 0. max weighted matching? specifies whether PARDISO uses a maximum weighted matching algorithm to permute large elements close to the diagonal. The default is FALSE. For highly indefinite symmetric matrices, National Instruments recommends setting max weighted matching? to TRUE. output MFlops on LU? specifies whether LabVIEW returns the number of floating-point operations in MFlops on factorization. Returning MFlops increases factorization time. The default is FALSE. memory option specifies whether LabVIEW uses the in-core or out-of-core (OOC) version of PARDISO. 0 Auto—Specifies that LabVIEW automatically determines the version of PARDISO to use. 1 In Core (default)—Specifies to use the in-core version of PARDISO where LabVIEW stores matrix factors in memory. 2 Out of Core—Specifies to use the out-of-core version of PARDISO where LabVIEW stores matrix factors in files on hard disk. Use this option for very large problems.
	solver ID returns the identification number associated with the PARDISO session this VI initializes.
	Diagnostic Info returns diagnostic information. Permutation returns the permutation vector that PARDISO actually used. LabVIEW returns the permutation vector only when output permutation? is TRUE. # of iterative refinement returns the number of iterative refinement that PARDISO actually performed. # of nonzeros in LU returns the number of nonzeros in the factorization. peak memory [KB] on symbolic LU returns the peak memory in kilobytes that PARDISO needed on symbolic factorization. allocated memory [KB] on symbolic LU returns the memory, in kilobytes, that PARDISO actually allocated on symbolic factorization. allocated memory [KB] on LU returns the memory, in kilobytes, that PARDISO actually allocated on numerical factorization. MFlops on LU returns the number of floating-point operations, in MFlops, on factorization. LabVIEW returns MFlops only when output MFlops on LU? is TRUE. # of positive eigenvalues returns the number of positive eigenvalues when the Input Matrix is symmetric. # of negative eigenvalues returns the number of negative eigenvalues when the Input Matrix is symmetric.
	error out contains error information. This output provides standard error out functionality.

PARDISO Advanced Initialization (SGL)

	Input Matrix specifies a square sparse matrix to analyze and perform symbolic factorization.
	matrix type specifies the matrix type of the Input Matrix. 0 General (default)—Specifies a square matrix. 1 Symmetric—Specifies a symmetric matrix. LabVIEW only uses the upper triangular part in the Input Matrix to solve the linear system. 2 Hermitian—Specifies a Hermitian matrix. This option is the same as Symmetric when the Input Matrix is a real-valued matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero. 3 Positive Definite—Specifies a symmetric positive definite or Hermitian positive definite matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero.
	error in describes error conditions that occur before this node runs. This input provides standard error in functionality.
	Solver Parameters specifies the set of parameters to use when configuring PARDISO. ordering method specifies the method of permutation in order to reduce fill-in on factorization. 0 Minimum Degree—Specifies to use the minimum degree algorithm. 1 Nested Dissection (default)—Specifies to use the nested dissection algorithm. 2 Parallel Nested Dissection—Specifies to use the parallel version of nested dissection algorithm. 3 User Defined Permutation—Specifies to use the User Defined Permutation as the permutation vector to reduce fill-in. User Defined Permutation specifies the permutation vector to reduce fill-in. LabVIEW uses the user defined permutation vector only when ordering method is User Defined Permutation. output permutation? specifies whether LabVIEW returns the permutation vector that PARDISO actually used. The default is FALSE. max # of iterative refinement specifies the maximum number of iterative refinement that PARDISO can perform. The default is 0. max weighted matching? specifies whether PARDISO uses a maximum weighted matching algorithm to permute large elements close to the diagonal. The default is FALSE. For highly indefinite symmetric matrices, National Instruments recommends setting max weighted matching? to TRUE. output MFlops on LU? specifies whether LabVIEW returns the number of floating-point operations in MFlops on factorization. Returning MFlops increases factorization time. The default is FALSE. memory option specifies whether LabVIEW uses the in-core or out-of-core (OOC) version of PARDISO. 0 Auto—Specifies that LabVIEW automatically determines the version of PARDISO to use. 1 In Core (default)—Specifies to use the in-core version of PARDISO where LabVIEW stores matrix factors in memory. 2 Out of Core—Specifies to use the out-of-core version of PARDISO where LabVIEW stores matrix factors in files on hard disk. Use this option for very large problems.
	solver ID returns the identification number associated with the PARDISO session this VI initializes.
	Diagnostic Info returns diagnostic information. Permutation returns the permutation vector that PARDISO actually used. LabVIEW returns the permutation vector only when output permutation? is TRUE. # of iterative refinement returns the number of iterative refinement that PARDISO actually performed. # of nonzeros in LU returns the number of nonzeros in the factorization. peak memory [KB] on symbolic LU returns the peak memory in kilobytes that PARDISO needed on symbolic factorization. allocated memory [KB] on symbolic LU returns the memory, in kilobytes, that PARDISO actually allocated on symbolic factorization. allocated memory [KB] on LU returns the memory, in kilobytes, that PARDISO actually allocated on numerical factorization. MFlops on LU returns the number of floating-point operations, in MFlops, on factorization. LabVIEW returns MFlops only when output MFlops on LU? is TRUE. # of positive eigenvalues returns the number of positive eigenvalues when the Input Matrix is symmetric. # of negative eigenvalues returns the number of negative eigenvalues when the Input Matrix is symmetric.
	error out contains error information. This output provides standard error out functionality.

PARDISO Advanced Initialization (CDB)

	Input Matrix specifies a square sparse matrix to analyze and perform symbolic factorization.
	matrix type specifies the matrix type of the Input Matrix. 0 General (default)—Specifies a square matrix. 1 Symmetric—Specifies a symmetric matrix. LabVIEW only uses the upper triangular part in the Input Matrix to solve the linear system. 2 Hermitian—Specifies a Hermitian matrix. This option is the same as Symmetric when the Input Matrix is a real-valued matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero. 3 Positive Definite—Specifies a symmetric positive definite or Hermitian positive definite matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero.
	error in describes error conditions that occur before this node runs. This input provides standard error in functionality.
	Solver Parameters specifies the set of parameters to use when configuring PARDISO. ordering method specifies the method of permutation in order to reduce fill-in on factorization. 0 Minimum Degree—Specifies to use the minimum degree algorithm. 1 Nested Dissection (default)—Specifies to use the nested dissection algorithm. 2 Parallel Nested Dissection—Specifies to use the parallel version of nested dissection algorithm. 3 User Defined Permutation—Specifies to use the User Defined Permutation as the permutation vector to reduce fill-in. User Defined Permutation specifies the permutation vector to reduce fill-in. LabVIEW uses the user defined permutation vector only when ordering method is User Defined Permutation. output permutation? specifies whether LabVIEW returns the permutation vector that PARDISO actually used. The default is FALSE. max # of iterative refinement specifies the maximum number of iterative refinement that PARDISO can perform. The default is 0. max weighted matching? specifies whether PARDISO uses a maximum weighted matching algorithm to permute large elements close to the diagonal. The default is FALSE. For highly indefinite symmetric matrices, National Instruments recommends setting max weighted matching? to TRUE. output MFlops on LU? specifies whether LabVIEW returns the number of floating-point operations in MFlops on factorization. Returning MFlops increases factorization time. The default is FALSE. memory option specifies whether LabVIEW uses the in-core or out-of-core (OOC) version of PARDISO. 0 Auto—Specifies that LabVIEW automatically determines the version of PARDISO to use. 1 In Core (default)—Specifies to use the in-core version of PARDISO where LabVIEW stores matrix factors in memory. 2 Out of Core—Specifies to use the out-of-core version of PARDISO where LabVIEW stores matrix factors in files on hard disk. Use this option for very large problems.
	solver ID returns the identification number associated with the PARDISO session this VI initializes.
	Diagnostic Info returns diagnostic information. Permutation returns the permutation vector that PARDISO actually used. LabVIEW returns the permutation vector only when output permutation? is TRUE. # of iterative refinement returns the number of iterative refinement that PARDISO actually performed. # of nonzeros in LU returns the number of nonzeros in the factorization. peak memory [KB] on symbolic LU returns the peak memory in kilobytes that PARDISO needed on symbolic factorization. allocated memory [KB] on symbolic LU returns the memory, in kilobytes, that PARDISO actually allocated on symbolic factorization. allocated memory [KB] on LU returns the memory, in kilobytes, that PARDISO actually allocated on numerical factorization. MFlops on LU returns the number of floating-point operations, in MFlops, on factorization. LabVIEW returns MFlops only when output MFlops on LU? is TRUE. # of positive eigenvalues returns the number of positive eigenvalues when the Input Matrix is symmetric. # of negative eigenvalues returns the number of negative eigenvalues when the Input Matrix is symmetric.
	error out contains error information. This output provides standard error out functionality.

PARDISO Advanced Initialization (CSG)

	Input Matrix specifies a square sparse matrix to analyze and perform symbolic factorization.
	matrix type specifies the matrix type of the Input Matrix. 0 General (default)—Specifies a square matrix. 1 Symmetric—Specifies a symmetric matrix. LabVIEW only uses the upper triangular part in the Input Matrix to solve the linear system. 2 Hermitian—Specifies a Hermitian matrix. This option is the same as Symmetric when the Input Matrix is a real-valued matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero. 3 Positive Definite—Specifies a symmetric positive definite or Hermitian positive definite matrix. LabVIEW only uses the upper triangular part of the Input Matrix to solve the linear system. LabVIEW also considers all diagonal elements in the Input Matrix to have an imaginary part of zero.
	error in describes error conditions that occur before this node runs. This input provides standard error in functionality.
	Solver Parameters specifies the set of parameters to use when configuring PARDISO. ordering method specifies the method of permutation in order to reduce fill-in on factorization. 0 Minimum Degree—Specifies to use the minimum degree algorithm. 1 Nested Dissection (default)—Specifies to use the nested dissection algorithm. 2 Parallel Nested Dissection—Specifies to use the parallel version of nested dissection algorithm. 3 User Defined Permutation—Specifies to use the User Defined Permutation as the permutation vector to reduce fill-in. User Defined Permutation specifies the permutation vector to reduce fill-in. LabVIEW uses the user defined permutation vector only when ordering method is User Defined Permutation. output permutation? specifies whether LabVIEW returns the permutation vector that PARDISO actually used. The default is FALSE. max # of iterative refinement specifies the maximum number of iterative refinement that PARDISO can perform. The default is 0. max weighted matching? specifies whether PARDISO uses a maximum weighted matching algorithm to permute large elements close to the diagonal. The default is FALSE. For highly indefinite symmetric matrices, National Instruments recommends setting max weighted matching? to TRUE. output MFlops on LU? specifies whether LabVIEW returns the number of floating-point operations in MFlops on factorization. Returning MFlops increases factorization time. The default is FALSE. memory option specifies whether LabVIEW uses the in-core or out-of-core (OOC) version of PARDISO. 0 Auto—Specifies that LabVIEW automatically determines the version of PARDISO to use. 1 In Core (default)—Specifies to use the in-core version of PARDISO where LabVIEW stores matrix factors in memory. 2 Out of Core—Specifies to use the out-of-core version of PARDISO where LabVIEW stores matrix factors in files on hard disk. Use this option for very large problems.
	solver ID returns the identification number associated with the PARDISO session this VI initializes.
	Diagnostic Info returns diagnostic information. Permutation returns the permutation vector that PARDISO actually used. LabVIEW returns the permutation vector only when output permutation? is TRUE. # of iterative refinement returns the number of iterative refinement that PARDISO actually performed. # of nonzeros in LU returns the number of nonzeros in the factorization. peak memory [KB] on symbolic LU returns the peak memory in kilobytes that PARDISO needed on symbolic factorization. allocated memory [KB] on symbolic LU returns the memory, in kilobytes, that PARDISO actually allocated on symbolic factorization. allocated memory [KB] on LU returns the memory, in kilobytes, that PARDISO actually allocated on numerical factorization. MFlops on LU returns the number of floating-point operations, in MFlops, on factorization. LabVIEW returns MFlops only when output MFlops on LU? is TRUE. # of positive eigenvalues returns the number of positive eigenvalues when the Input Matrix is symmetric. # of negative eigenvalues returns the number of negative eigenvalues when the Input Matrix is symmetric.
	error out contains error information. This output provides standard error out functionality.

PARDISO Advanced Initialization Details

The following table lists the support characteristics of this VI.

Supported on RT targets	Yes
Suitable for bounded execution times on RT	No