On Thu, Jun 30, 2016 at 8:48 PM, Jeff Hammond <[email protected]> wrote:

...

The default settings often don’t work well for 3D problems.

Are 2D (1D?) problems really the common case for PDE solvers? Aren't interesting problems 3D? Shouldn't the defaults be set to optimize for 3D?

Lots of people do 2D in production (it might just be part of a larger 3D algorithm like the XGC code). Also 2D is useful for debugging, regression, incremental development. I would prefer that hypre's defaults be for 3D and I like Barry's suggestion of changing the default with the dimension. We can get a shot at the dimension, without any API changes, by just inferring from the block size. Better than nothing. But I do find it annoying to search around for my 3D hypre parameters when I'm (customers) doing a 3D problems. I am sure lots of cycles are wasted by PETSc users using hypre in 3D with the defaults.

Jeff

-- Jeff Hammond [email protected] http://jeffhammond.github.io/