• UF Sparse Matrix Collection
  • Matrix group: Koutsovasilis
  • Click here for a description of the Koutsovasilis group.
  • Click here for a list of all matrices
  • Click here for a list of all matrix groups


  • Matrix: Koutsovasilis/F2
  • Description: AUDI engine piston rod; Panos Koutsovasilis, Technische Univ. Dresden
  • download as a MATLAB mat-file, file size: 37 MB. Use UFget(1861) or UFget('Koutsovasilis/F2') in MATLAB.
  • download in Matrix Market format
  • download in Rutherford/Boeing format

    Koutsovasilis/F2

    Koutsovasilis/F2 graph

    Matrix properties
    number of rows71,505
    number of columns71,505
    nonzeros5,294,285
    structural full rank?yes
    structural rank71,505
    # of blocks from dmperm1
    # strongly connected comp.1
    entries not in dmperm blocks0
    explicit zero entries0
    nonzero pattern symmetrysymmetric
    numeric value symmetrysymmetric
    typereal
    structuresymmetric
    Cholesky candidate?yes
    positive definite?no

    authorP. Koutsovasilis
    editorT. Davis
    date2007
    kindstructural problem
    2D/3D problem?yes

    Notes:

    Stiffness matrix from an AUDI engine piston rod.  This matrix is an ill-     
    condition symmetric indefinite matrix.  In MATLAB 7.4, the matrix is         
    factorized three times in x=A\b.  Inside backslash, CHOLMOD is tried first   
    since the matrix is symmetric and all diagonal entries are positive.  CHOLMOD
    fails since the matrix is indefinite.  Next, UMFPACK is used with default    
    pivot tolerances which maintain sparsity at the (rare) expense of a slight   
    decrease in accuracy.  This succeeds, but the result is flagged as           
    potentially inaccurate because the condition estimate is high.  UMFPACK is   
    then used again with more conservative tolerances (but more fill-in).        
    

    Ordering statistics:AMD METIS
    nnz(chol(P*(A+A'+s*I)*P'))20,756,208 21,041,811
    Cholesky flop count1.2e+10 1.2e+10
    nnz(L+U), no partial pivoting41,440,911 42,012,117
    nnz(V) for QR, upper bound nnz(L) for LU40,210,857 40,468,056
    nnz(R) for QR, upper bound nnz(U) for LU109,796,898 107,281,795

    Maintained by Tim Davis, last updated 04-May-2008.
    Matrix pictures by cspy, a MATLAB function in the CSparse package.
    Matrix graphs by Yifan Hu, AT&T Labs Visualization Group.