[tuhopuu-devel] Help for tuhopuu's uv mapping - continued (2)

[Brecht Van Lommel]

Hi,

On Thu, 2004-05-13 at 09:50, Bruno Levy wrote:
> Hi there,
> I've just included support for SuperLU 3.0 in 
> http://www.loria.fr/~levy/OpenNL,
> it is worth it, here is the output of the included test program
> (computes LSCM on a .obj file):
> 
> ---> without SuperLU
> Loading ...
> Loaded 12917 vertices and 12812 facets
> Using Jacobi pre-conditioned conjugate gradient
> Solving ...
> Solver time: 15.74
> Saving ...
> 
> ---> with SuperLU
> Loading ...
> Loaded 12917 vertices and 12812 facets
> Using SuperLU, cool !
> Solving ...
> Solver time: 2.23
> Saving ...
> 
> Note1: if you use SuperLU directly, make sure that you use
> the COLAMD column permutation, there is a *dramatic*
> impact on performances:
>    on this example, without COLAMD -> 60 seconds
>    with COLAMD -> 2.23 seconds
> 
> Note2: SuperLU 3.0 support symmetric storage, this halves
>   storage requirements, I'll try to use it (since SuperLU is
>   a memory-eater, this can be worth it).
> 
> -- Bruno.
> 

Thanks a lot for all the efforts you are doing! We will be using OpenNL
for solving the lineair system, and I'll be commiting it to tuhopuu cvs
in one of the following days. 

We decided to do the LSCM part in blender's code, with blender's data
structures, but the main.cpp with LSCM implementation will certainly
help us implementing it.

Symmetric storage indeed seems to be an interesting thing to
investigate.

Thanks again for all the help!

Cheers,
Brecht