ckees - 2005-06-21


Could someone point me to more details on the jacobi preconditioner in PySparse?  For a linear system A*x = b, the usual formulation of damped jacobi is

x_1 = x_0 + \omega*D^{-1} (b - A*x)

When I run
>>>preconA = jacobi(A,omega,ns)

I get a different answer than with a simple implementation of Jacobi iteration. Does precon just repeatedly multiply by D^{-1}?