[GRASS-dev] Parallel computing for r.sun

[Hamish]

Hi Jan,

> So, what do you think about r.los? Is there any
> implementation of the parallel computing?
> 
> If there is r.los also available for parallel computing we
> will try v.surf.bspline or v.surf.bspline, that are in the
> area of our interest as well.

the current r.los implementation does not scale at all well to
large array sizes. So r.viewshed is probably the better of
those two to look at, the only thing to consider is that it is
written in C++, if that matters to you.

see
  http://grasswiki.osgeo.org/wiki/GPU#Modules_of_interest_to_be_parallelized
and
  http://grasswiki.osgeo.org/wiki/OpenMP#Candidates

"r.los (to be replaced by r.viewshed after last few bugs are fixed)"

(and the other pages in the Parallelization wiki category)



have fun,
Hamish