About these classes:

CRS.h

• This file contains class to store a sparse matrix in Compressed Row Storage format
• It can also be used independently to store any sparse matrix in CRS format
• Creating objects: CRS<data-type-of-elements> obj-name(input-matrix, flag)
• Input matrix: A 2-D vector. vector<vector<data-type> >
• flag = 0, if you want to delete the input matrix after storing in CRS format. (Recommended for large Matrices)
  = 1, if you don't want to delete input matrix
• Currently Methods available includes just Matrix-Vector multiplication. Usage: C=A.MVMult(B); // performs c=A.b where, A=CRS object created with the input matrix
  B=Vector to be multiplied
  C=Result Vector
  

CGCRYLOV.h

• This file contains class to solve AX=B system
• Creating objects: CGCRYLOV<data-type-of-elements>=obj-name(problem-size)
• Requires CRS.h to be in the same folder
• Require CBLAS
• Methods available:

  • X=Solve(A,B)

    Where,
    A=2D vector vector<vector<data-type> >
    B=1D vector
    X=Result vector
    This method solves any AX=B for positive definite A matrix.

  • X=SolvePreCond(A,M,B)

    Where,
    A=2D vector vector<vector<data-type> >
    M=2D vector vector<vector<data-type> >
    B=1D vector
    X=Result vector
    This method solves any AX=B for positive definite A matrix with preconditioner M.

  • X=SolveCRS(A,M,B)

    Where,
    A=2D vector vector<vector<data-type> >
    B=1D vector
    X=Result vector
    This method solves any AX=B for positive definite sparse A matrix.

  • X=SolvePreCondCRS(A,M,B)

    Where,
    A=2D vector vector<vector<data-type> >
    M=2D vector vector<vector<data-type> >
    B=1D vector
    X=Result vector
    This method solves any AX=B for positive definite sparse A matrix with preconditioner M.