Blender Git Loki

Render Branch: implicit solver for hair now calculates forces and does
integration separate per hair. This increase the stability of the hair.

I found that usually the conjugate gradient takes about 4-8 steps, but
it sometimes takes 20 or even does not converge, and returns bogus
velocities. I'm not entirely sure why that is, but it seems logical
that a smaller system would be more numerically stable. What I suspect
is that since there are many hairs, and it tests for the average error
of those hairs, a few hairs may be able to hide their high error in this
average.

I've tried to not make the code too ugly but's there's a few stupid
hacks, maybe eventually it would be better to have a separate hair and
cloth function that calls common functions.

Also for threading, hairs can now be distributed over threads. There
still seems to be too much thread overhead though, should find a way
to keep all threads running for the entire timestep.