Interface [ITL Home] Programmers Guide
  Contents | Index |  Search 


Category: itl Component type:concept
Description
An ITL Interface provides the basic linear operations required by iterative solvers.
  • Conjugate Gradient(cg)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot_conj(x, y); Conjugate inner product: x * yT
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • Conjugate Gradient Squared (cgs)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot(x, y); Inner product: x * y
    add(x, y); y += x
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • BiConjugate Gradient (bicg)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    trans_mult(A, x, y);y = AT * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    trans_solve(M, x, y);Solving tranpose preconditioner system
    dot(x, y); Inner product: x * y
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • BiConjugate Gradient Stabilized (bicgstab)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot(x, y); Inner product: x * y
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • Chebyshev Iteration (cheby)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • Richardson Iteration (richardson)
    mult(A, x, y, z);z = y + A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    add(x, y); y += x
    two_norm(x);The two norm of x
  • Generalized Conjugate Residual (gcr)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot_conj(x, y); Conjugate inner product: x * yT
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
    internal_matrix_traits::MatrixInternal matrix type
  • Generalized Minimal Residual (gmres)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    scale(x, alpha); Scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot_conj(x, y); Conjugate inner product: x * yT
    add(x, y); y += x
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
    internal_matrix_traits::MatrixInternal matrix type
    upper_tri_solve(A, x, i);Backward substitution for upper triangular part of Hessenberg matrix
    Orthogonalizer;To perform orthogonalization
  • Quasi-Minimal Residual Without Lookahead (qmr)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    trans_mult(A, x, y);y = AT * x
    scaled(x, alpha);Delaying scaling vector x
    scale(x, alpha); Scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    trans_solve(M, x, y);Solving tranpose preconditioner system
    dot(x, y); Inner product: x * y
    add(x, y); y += x
    add(x, y, z); z = x + y
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
  • Transpose Free Quasi-Minimal Residual Without Lookahead (tfqmr)
    mult(A, x, y, z);z = y + A * x
    mult(A, x, y);y = A * x
    scaled(x, alpha);Delaying scaling vector x
    size(x);The dimension of vector space
    solve(M, x, y);Solving preconditioner system
    dot(x, y); Inner product: x * y
    add(x, y, z); z = x + y
    add(x, y, z, r); r = x + y + z
    copy(x, y);Copying elements of x to resulting vector y
    two_norm(x);The two norm of x
Models
Notes
See also

[ITL Home] Copyright © 1998,1999 University of Notre Dame. All Rights Reserved.