Indeed, I have used Linear Variational Solver only for eta^{n+1} and phi^{n+1} (and q^{n+1/2} ). Phi^{n+1/2} is still obtained from the NL variational solver. The new log summary is attached. Thanks, Floriane ________________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : lundi 14 mars 2016 16:12 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] Debug and Optimised mode On 14/03/16 14:22, Floriane Gidel [RPG] wrote:
Thanks a lot Lawrence! I'll try to reformulate using linear solvers and I'll let you know how it goes. My domain has the shape of the domain attached, where the horizontal side has to be 100-300m (while the wave length is of order 1m which requires a mesh refinement of order 0.1 max). I tried to run the simulations on a faster machine but it did not improve much. But it might indeed after I linearise the solvers. I'll come back to you when it's done.
Please note, looking again I see that at least the equation to compute phi^{n+1/2} *is* nonlinear, so please check my working! Lawrence