Interesting, thanks.  I didn't expect the FFC:COFFEE ratio to be so form-dependent.

On 17 September 2015 at 16:45, Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> wrote:

> On 17 Sep 2015, at 12:14, Andrew McRae <A.T.T.McRae@bath.ac.uk> wrote:
>
> COFFEE shouldn't be particularly slow unless you forgot to update in the last 3 months...

My timings suggested almost all the time is spent in coffee (O2)

Lawrence
_______________________________________________
firedrake mailing list
firedrake@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/firedrake