Real-Time Physics Simulation Forum

Got some free time recently, and implemented NNCG . It is quite good, I must say. As expected, it is a little bit worse in terms of pure linear system convergence (pasting as URL since I couldn't make the BBcode width to work here): https://raw.githubusercontent.com/avoroshilov/physics_playground/ma...

Btw, Erwin, you mentioned very high iteration counts - that depends on the definition of very high. At the scene of box stacking there is clear benefit of our solver over the PGS at 30 iterations, and generally it depends on the scene complexity: the more complex the scene, the more iteration our so...

Dirk Gregorius , Erwin Coumans Glad to share! I remember Dirk asked about the NNCG last time we shared our demos, but I didn't have time to implement NNCG back then, or even look too closely. Now, I have that on the list of things to do, and if I'll get to it, I will share the results. From the ver...

Hello! Several years ago we already shared this kind of solver, but last time our demonstration framework was not very compelling (for example, it was lacking collision detection and hence only very basic scenes with constraint force limits could be tested) - so I decided to revisit it now as I've g...