[Bf-committers] SPH update (Farsthary)

[Raul Fernandez Hernandez]

Hi all :)

  Thanks to Jahka´s very helpfull suggestions I could use the KDtree
acceleration structure in the SPH particles, not without issues (a non
optimal parameters set could lead to slower simulations (slower than the
brute force SPH method) ) that caused me confusion at first time, but
choosing ¨good¨ parameters inmediately shows the advantage of KD tree
implementation.
   (Could be usefull implement in the future some presets)
  I have run some test here with 10k particles at relatively fast speeds
(where the raw implementation runs ok up to 4k-5k particles acording to
hardware)

  I have sended him now the code, for review and his master touch,
currently is not suited for public release since I haven´t implemented
the full UI for it  (I definately have a problem with GUI and RNA
programming :S) and also is subject to some design considerations
acording to the physical/logical meaning.

  One killer feature of the SPH fluids will be the fluid-fluid
interactions (multifluid simulation), though currently we aim to polish
the single fluid simulation the code has the potential to be extended
relatively easy to multifluids (the KDtree make it trivial). Once the
single fluid SPH is production ready then we could go for Multifluids.

                                 Cheers to all :)