dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

Jattebra!

Jag kan se om jag kan hitta nagot ocksa. Haller precis pa med en ny
grej i FFC (som forhoppningsvis kommer att snabba upp kompileringen
med FFC rejalt) men jag borde fa tid senare i veckan att hitta felet.

Har du nagot bra testproblem? Det med 20 komponenter?

/Anders

On Sun, Nov 20, 2005 at 04:19:15PM +0100, dolfin@xxxxxxxxxx wrote:
> Commit from johanjan (2005-11-20 16:19 CET)
> --------------------
> 
> ODE:
> 
> Fixed bug in mcG(q). Previously, f at the right-hand endpoint was
> saved even if the slab was rejected. This is now fixed.
> 
> The error for mcG(1) in the benchmark has been reduced by up to one
> magnitude (runtime is unchanged). The error is still larger than for
> cG(1) though, so there appears to remain at least one source of
> error. We'll have to keep looking a bit more.
> 
>   dolfin  src/demo/solvers/ode/reaction/bench.log        1.12
>   dolfin  src/kernel/ode/MultiAdaptiveTimeSlab.cpp       1.46
>   dolfin  src/kernel/ode/dolfin/MultiAdaptiveTimeSlab.h  1.24
> 
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> 

-- 
Anders Logg
Research Assistant Professor
Toyota Technological Institute at Chicago
http://www.tti-c.org/logg/