Package jdistlib.accelerator

Class SparseCholeskyFactor

java.lang.Object
jdistlib.accelerator.SparseCholeskyFactor
public final class SparseCholeskyFactor extends Object
Immutable FP64 sparse Cholesky factor with reusable solves.

  • Method Summary

    Modifier and Type
    Method
    Description
    int
    dimension()
     
    double
    logDeterminant()
     
    CsrMatrix
    lower()
    Returns the lower factor in permuted coordinates and one-based CSR storage.
    int
    nonzeroCount()
     
    int[]
    permutation()
    Returns the new-to-original symmetric permutation used by the factor.
    double[]
    solve(double[] right)
    Solves A*x=right in original input coordinates.
    double[]
    solve(double[] right, int columns)
    Solves a row-major dimension-by-columns right side in original coordinates.
    void
    solveInPlace(double[] right, int columns)
    Replaces a row-major dimension-by-columns right side with its solution.

    Methods inherited from class java.lang.Object

    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

  • Method Details

    • dimension

      public int dimension()

    • nonzeroCount

      public int nonzeroCount()

    • logDeterminant

      public double logDeterminant()

    • permutation

      public int[] permutation()
      Returns the new-to-original symmetric permutation used by the factor.

    • lower

      public CsrMatrix lower()
      Returns the lower factor in permuted coordinates and one-based CSR storage.

    • solve

      public double[] solve(double[] right)
      Solves A*x=right in original input coordinates.

    • solve

      public double[] solve(double[] right, int columns)
      Solves a row-major dimension-by-columns right side in original coordinates.

    • solveInPlace

      public void solveInPlace(double[] right, int columns)
      Replaces a row-major dimension-by-columns right side with its solution.