ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Wed, Oct 05, 2005 at 05:43:14PM +0200, Garth N. Wells wrote:

> > > I can't test big problems because BLAS mode has *huge* memory
> > > requirements when using higher order elements. For Poisson equation
> > > with p = 10, I can't use any more than 35x35 elements without having
> > > a PETSc memory error.
> > 
> > p = 10 is pretty high... you can add a line
> 
> If I use p =3 on triangles, I can't use more than 120k elements - it requires
> more than 1GB of memory. Without BLAS, assembly is no problem. Also, with the
> p=10 example I was referring to triangles.

1GB with p = 3 on triangles for Poisson, and only in BLAS mode?

That sounds like a memory leak. BLAS mode should only be a little
overhead in memory (if any). The only difference is that data is
stored in arrays rather than in the code itself which should be
about the same total memory usage. 

> > > Separate issue - with more complicated forms, I have trouble with anything
> > > around p > 2 due to the memory requirements of ffc. Are there any simple
> > > solutions to this?
> 
> With Poisson, I can go really high. I'll send the Cahn-Hilliard form shortly to
> try out. With Cahn-Hilliard I'm quite limited.

ok.

/Anders