Re: [petsc-dev] [petsc-users] Poor weak scaling when solving successive linearsystems

[Mark Adams]

>
>
> Just -pc_type hypre instead of -pc_type gamg.
>
>
And you need to have configured PETSc with hypre.