[petsc-users] chelosky vs multigrid

[Jed Brown]

Fazlul Huq via petsc-users <petsc-users at mcs.anl.gov> writes:

> Hello PETSc Developers,
>
> I am trying to solve a poisson equation using "-pc_type cholesky" and
> "-pc_type hypre -pc_hypre_type boomeramg" and I got that cholesky
> decomposition method takes
> less time then multigrid method.
> Is this expected? 

For small problems, yes.  If using Cholesky in practice, you may want to
install SuiteSparse or MUMPS and use -pc_factor_mat_solver_type cholmod
(or mumps).

> I thought that multigrid method solve faster than cholesky
> decomposition method.

For sufficiently large problems (especially in 3D), yes.