講師：　Prof. Mark Sussman (Florida State University)

Standard multigrid smoothers, e.g. jacobi method or symmetric gauss seidel method, perform well when multigrid is used as a preconditioner for the conjugate gradient method.  Unfortunately, performance can degrade under the following extenuating factors: discontinuous coefficients along a complex interface, computing on parallel platforms, computing with the aid of GPU acceleration, and computing on an underlying grid that has a large aspect ratio.  In this presentation, the benefits and drawbacks of the ``ILU smoother'', ``GPU accelerated jacobi smoother'', ``line relaxation smoother'', and ``red-black gauss-seidel smoother'' are studied.