Class SemiringCsrMatMult

java.lang.Object
org.flag4j.linalg.ops.sparse.csr.semiring_ops.SemiringCsrMatMult
public final class SemiringCsrMatMult extends Object
Utility class for computing matrix multiplication of a sparse CSR Semiring matrix.

  • Method Summary

    Modifier and Type
    Method
    Description
    static <T extends Semiring<T>>
    void
    standard(Shape shape1, T[] src1Entries, int[] src1RowPointers, int[] src1ColIndices, Shape shape2, T[] src2Entries, int[] src2RowPointers, int[] src2ColIndices, T[] destEntries, Semiring<T> zero)
    Computes the matrix multiplication between two sparse CSR matrices.
    static <T extends Semiring<T>>
    SparseMatrixData<T>
    standardToSparse(Shape shape1, T[] src1Entries, int[] src1RowPointers, int[] src1ColIndices, Shape shape2, T[] src2Entries, int[] src2RowPointers, int[] src2ColIndices)
    Computes the matrix multiplication between two sparse CSR matrices and returns the result as a sparse matrix.
    static <T extends Semiring<T>>
    void
    standardVector(Shape shape, T[] src1, int[] rowPointers, int[] colIndices, int size, T[] src2, int[] indices, T[] dest, T zero)
    Computes the matrix-vector multiplication between a sparse CSR matrix and a sparse COO vector.

    Methods inherited from class java.lang.Object

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

  • Method Details

    • standard

      public static <T extends Semiring<T>> void standard(Shape shape1, T[] src1Entries, int[] src1RowPointers, int[] src1ColIndices, Shape shape2, T[] src2Entries, int[] src2RowPointers, int[] src2ColIndices, T[] destEntries, Semiring<T> zero)
      Computes the matrix multiplication between two sparse CSR matrices. The result is a dense matrix.
      Parameters:
      shape1 - Shape of the first matrix in the matrix multiplication.
      src1Entries - Non-zero data of the first CSR matrix in the matrix multiplication.
      src1RowPointers - Non-zero row pointers of the first CSR matrix in the matrix multiplication.
      src1ColIndices - Non-zero column indices of the first CSR matrix in the matrix multiplication.
      shape2 - Shape of the second matrix in the matrix multiplication.
      src2Entries - Non-zero data of the second CSR matrix in the matrix multiplication.
      src2RowPointers - Non-zero row pointers of the second CSR matrix in the matrix multiplication.
      src2ColIndices - Non-zero column indices of the second CSR matrix in the matrix multiplication.
      destEntries - Array to store the dense result of the matrix multiplication in (modified).
      zero - The zero value of the semiring.

    • standardToSparse

      public static <T extends Semiring<T>> SparseMatrixData<T> standardToSparse(Shape shape1, T[] src1Entries, int[] src1RowPointers, int[] src1ColIndices, Shape shape2, T[] src2Entries, int[] src2RowPointers, int[] src2ColIndices)

      Computes the matrix multiplication between two sparse CSR matrices and returns the result as a sparse matrix.

      Warning: This method may be slower than standard(Shape, Semiring[], int[], int[], Shape, Semiring[], int[], int[], Semiring[], Semiring) if the result of multiplying this matrix with src2 is not very sparse. Further, multiplying two sparse matrices (even very sparse matrices) may result in a dense matrix so this method should be used with caution.

      Parameters:
      shape1 - Shape of the first matrix in the matrix multiplication.
      src1Entries - Non-zero data of the first CSR matrix in the matrix multiplication.
      src1RowPointers - Non-zero row pointers of the first CSR matrix in the matrix multiplication.
      src1ColIndices - Non-zero column indices of the first CSR matrix in the matrix multiplication.
      shape2 - Shape of the second matrix in the matrix multiplication.
      src2Entries - Non-zero data of the second CSR matrix in the matrix multiplication.
      src2RowPointers - Non-zero row pointers of the second CSR matrix in the matrix multiplication.
      src2ColIndices - Non-zero column indices of the second CSR matrix in the matrix multiplication.
      Returns:
      Sparse CSR matrix resulting from the matrix multiplication of the two sparse CSR matrices.

    • standardVector

      public static <T extends Semiring<T>> void standardVector(Shape shape, T[] src1, int[] rowPointers, int[] colIndices, int size, T[] src2, int[] indices, T[] dest, T zero)
      Computes the matrix-vector multiplication between a sparse CSR matrix and a sparse COO vector.
      Parameters:
      shape - Shape of the CSR matrix.
      src1 - Non-zero data of the CSR matrix
      rowPointers - Non-zero row pointers of the CSR matrix.
      colIndices - Non-zero column indices of the CSR matrix.
      size - Full size of the COO vector.
      src2 - Non-zero data of the COO vector.
      indices - Non-zero indices of the COO Vector.
      dest - Array to store the dense result of the matrix-vector multiplication in.
      zero - Zero element of the semiring. If null then the zero value will attempt to be discerned from src1. However, if zero == null && src1.length == 0, then zeros in the resulting dense vector will instead be null.
      Throws:
      IllegalArgumentException - If shape.get(1) != size or dest.length != size.