sparsekit_test

sparsekit_test, a Fortran90 code which calls sparsekit(), which carries out a number of operations on sparse matrices, particularly conversion between various sparse formats.

Licensing:

The information on this web page is distributed under the MIT license.

Related Data and Programs:

sparsekit, a Fortran90 code which carries out a number of operations on sparse matrices, particularly conversion between various sparse formats.

Source Code:

• sparsekit_test.sh, runs all the tests.

Sample problem 1:

• sparsekit_test01.f90, a sample calling program.
• sparsekit_test01.txt, the output file.

Sample problem 2:

• sparsekit_test02.f90, a sample calling program.
• sparsekit_test02.txt, the output file.

Sample problem 3:

• sparsekit_test03.f90, a sample calling program.
• sparsekit_test03.txt, the output file.

Sample problem 4 takes a banded matrix of order 16, stored as a dense matrix, converts it to CSR format and sorted CSR format.

• sparsekit_test04.f90, a sample calling program.
• sparsekit_test04.txt, the output file.

Sample problem 5:

• sparsekit_test05.f90, a sample calling program.
• sparsekit_test05.txt, the output file.

Sample problem 6:

• sparsekit_test06.f90, a sample calling program.
• sparsekit_test06.txt, the output file.

Sample problem 7:

• sparsekit_test07.f90, a sample calling program.
• sparsekit_test07.txt, the output file.

Sample problem 8 generates three sample matrices from the Zlatev set, and writes them to Harwell-Boeing format files:

• sparsekit_test08.f90, a sample calling program.
• sparsekit_test08.txt, the output file.
• zlatev1_hb.txt, Zlatev sample matrix 1.
• zlatev2_hb.txt, Zlatev sample matrix 2.
• zlatev3_hb.txt, Zlatev sample matrix 3.

Sample problem 9:

• sparsekit_test09.f90, a sample calling program.
• sparsekit_test09.txt, the output file.

Sample problem 10:

• sparsekit_test10.f90, a sample calling program.
• sparsekit_test10.txt, the output file.

Sample problem 11:

• sparsekit_test11.f90, a sample calling program.
• sparsekit_test11.txt, the output file.

Sample problem 12:

• sparsekit_test12.f90, a sample calling program.
• sparsekit_test12.txt, the output file.

Sample problem 13:

• sparsekit_test13.f90, a sample calling program.
• sparsekit_test13.txt, the output file.

Sample problem 14 generates a sample CSR matrix and converts it to an NCF (nonsymmetric coordinate format) used by NSPCG.

• sparsekit_test14.f90, a sample calling program.
• sparsekit_test14.txt, the output file.

List of Routines:

• AMASK extracts a sparse matrix from a masked input matrix.
• AMUB performs the matrix by matrix product C = A * B.
• AMUBDG gets the number of nonzero elements in each row of A*B.
• AMUDIA performs the matrix by matrix product B = A * Diag (in place)
• AMUX multiplies a CSR matrix A times a vector.
• AMUXD multiplies a DIA matrix times a vector.
• AMUXE multiplies an ELL matrix times a vector.
• AMUXJ multiplies a JAD matrix times a vector.
• APLB performs the CSR matrix sum C = A + B.
• APLB1 performs the matrix sum C = A+B for matrices in sorted CSR format.
• APLBDG gets the number of nonzero elements in each row of A+B and the total
• APLDIA adds a diagonal matrix to a general sparse matrix: B = A + Diag
• APLSB performs the matrix linear combination C = A+s*B
• APLSB1 performs the operation C = A+s B for matrices in sorted CSR format.
• APLSBT performs the matrix sum C = A + B'.
• APLSCA adds a scalar to the diagonal entries of a sparse matrix A :=A + s I
• APMBT performs the matrix sum C = A + B' or C = A - B'.
• ASSMB1 ???
• ASSMBO ???
• ATMUX computes A' * x for a CSR matrix A.
• BLKCHK checks whether the input matrix is a block
• BLKFND attemptps to determine whether or not the input
• BNDCSR converts Banded Linpack format to Compressed Sparse Row format.
• BOUND counts the number of boundary points and
• BSORT2 simple bubble sort for getting the ncut largest
• BSRCSR converts Block Sparse Row to Compressed Sparse Row.
• BSTEN calculates the correct block-stencil values for
• CHECKREF returns the expected the new number of nodes and
• CHKELMT checks the labeling within each element and reorders
• CNRMS gets the norms of each column of A. (choice of three norms)
• COICSR converts COO to CSR in place.
• COOCSR converts COO to CSR.
• COOELL converts coordinate format to ellpack format.
• COPMAT copies the matrix a, ja, ia, into the matrix ao, jao, iao.
• CPERM permutes the columns of a matrix.
• CSCAL scales the columns of A such that their norms are one on return
• CSORT sorts the elements of a matrix (stored in Compressed
• CSRBND converts Compressed Sparse Row to Banded (Linpack ) format.
• CSRBSR converts Compressed Sparse Row to Block Sparse Row
• CSRCOO converts Compressed Sparse Row to Coordinate
• CSRCSC converts Compressed Sparse Row to Compressed Sparse Column
• CSRDIA converts Compressed sparse row to diagonal format
• CSRDNS converts Compressed Sparse Row to Dense
• CSRELL converts Compressed Sparse Row to Ellpack - Itpack format
• CSRJAD converts Compressed Sparse Row to JAgged Diagonal storage.
• CSRLNK converts Compressed Sparse Row to Linked storage format.
• CSRMSR converts Compressed Sparse Row to Modified - Sparse Row
• CSRSSK converts Compressed Sparse Row to Symmetric Skyline Format
• CSRSSR converts Compressed Sparse Row to Symmetric Sparse Row
• DCN generates sparse square matrices of type D(N,C).
• DCSORT computes a permutation which, when applied to the
• DIACSR converts diagonal format to compressed sparse row
• DIAMUA performs the matrix by matrix product B = Diag * A (in place)
• DIAPOS returns the positions of the diagonal elements of a
• DINFO1 computes and prints matrix statistics.
• DIRIC takes into account the boundary conditions
• DLAUNY is a simple, nonoptimal Delaunay triangulation code.
• DNSCSR converts Dense to Compressed Row Sparse
• DPERM permutes the rows and columns of a matrix stored in CSR
• DSCALDG scales rows by diag where diag is either given (job=0)
• DUMP writes the matrix in a file, one row at a time in a nice readable
• DVPERM performs an in-place permutation of a real vector x
• ECN generates sparse (square) matrices of the type E(N,C).
• ELLCSR converts Ellpack-Itpack to Compressed Sparse Row.
• ESTIF3 constructs the element stiffness matrix using 3-node triangular elements
• EXPHES computes the Arnoldi basis and the corresponding
• EXPPRO computes an approximation to the vector
• EXPPROD computes an approximation to the vector
• EXTBDG extracts the main diagonal blocks of a
• FILTER removes any elements whose absolute value
• GEN57BL computes the sparse matrix in compressed
• GEN57PT computes the sparse matrix in compressed
• GENFEA generates finite element matrices for heat
• GENFEU generates finite element matrices for heat
• GETBWD gets the bandwidth of lower part and upper part of A.
• GETDIA extracts a given diagonal from a matrix stored in CSR
• GETELM returns the element a(i,j) of a matrix A,
• GETL extracts the lower triangular part of a matrix
• GETSTEN calculates the correct stencil values for
• GETU extracts the upper triangular part of a matrix
• GRADI3 constructs the first derivative of the shape functions.
• HES computes exp ( H dt) * y (1)
• HSOURC generates the load vector f in assembled form from the
• ILU0 is an ILU(0) preconditioner.
• ILUT is an ILUT preconditioner.
• INFDIA obtains information on the diagonals of A.
• IVPERM performs an in-place permutation of an integer vector
• JADSCR converts Jagged Diagonal Storage to Compressed Sparse Row
• LDSOL solves L x = y L = triangular. MSR format
• LDSOLC solves L x = y ; L = nonunit Low. Triang. MSC format
• LDSOLL solves L x = y L = triangular. Uses LEVEL SCHEDULING/MSR format
• LEVELS gets the level structure of a lower triangular matrix
• LNKCSR converts linked list storage format to Compressed Sparse Row format.
• LSOL solves L x = y ; L = lower unit triang. / CSR format
• LSOLC solves L x = y ; where L = unit lower trang. CSC format
• LUSOL0 performs a forward followed by a backward solve
• MARKGEN is a matrix generator for a markov model of a random walk on a triang. grid
• MATRF2 generates sparse (rectangular or square) matrices.
• MGSR is a modified gram - schmidt with partial reortho.
• MILU0 is a simple milu(0) preconditioner. *** *
• MSRCSR converts Modified - Sparse Row to Compressed Sparse Row
• PGMRES is an ILUT - Preconditioned GMRES solver.
• PLTMT creates a 'pic' file for plotting the pattern of
• PLTMTPS creates a 'PS' file for plotting the pattern of
• PRTMT writes a matrix in Harwell-Boeing format into a file.
• READMT reads a boeing/harwell matrix. handles right hand
• REFALL refines a finite element grid using triangular elements.
• RETMX returns in dd(*) the max absolute value of elements in row *.
• RNRMS gets the norms of each row of A. (choice of three norms)
• RPERM permutes the rows of a matrix in CSR format.
• RSCAL scales the rows of A such that their norms are one on return
• SSKSSR converts Symmetric Skyline Format to Symmetric Sparse Row format.
• SSRCSR converts Symmetric Sparse Row to (regular) Compressed Sparse Row
• SUBMAT extracts the submatrix A(i1:i2,j1:j2) and puts the result in
• TRANSP carries out in-place transposition routine.
• UDSOL solves U x = y ; U = upper triangular in MSR format
• UDSOLC Solves U x = y ; U = nonunit Up. Triang. MSC format
• UNASSBL ???
• USOL solves U x = y U = unit upper triangular.
• USOLC solves U x = y ; where U = unit upper trang. CSC format
• SAXPY CONSTANT TIMES A VECTOR PLUS A VECTOR.
• SDOT FORMS THE DOT PRODUCT OF TWO VECTORS.

You can go up one level to the Fortran90 source codes.