[ODE] Speeding up iterative techniques

[Russ Smith]

> So what exactly would you lose by using a non-LCP solver?  As far as I can
> tell, it's coulomb-ish friction (0 and infinite friction still works, as
> would slip maybe? -- not looking at the source right now) and fmax control
> (fmax is always infinity).  Am I missing anything?  How many people would
> be willing to give those up in the interest of speed?

LCP gives you: hard constraint-based contacts, finite friction (not 0 or
infinity), and 'velocity+fmax' motors. the contacts are the most
important one.

russ.

--
Russ Smith
http://www.q12.org/