Island management is a fundamental low level feature of physics engines and can have a big impact on solver design and performance. This was one of the first problems I decided to work on for Box2D version 3 (v3). Since I began working on v3 I’ve been comparing several algorithms for island management. My goal has been to make island building scale better with multiple CPU cores. Here are the three approaches I’ve considered:| box2d.org
Over the holidays I started working on version 3.0 of Box2D. I started closing out some issues for v2.4 and then moved onto a backlog task: speculative collision. Speculative collision is a technique for continuous collision that promises to remove the serial bottleneck of time of impact sub-stepping. I made good progress in my research and then I realized that this would be a major update to Box2D. And if I'm doing a major update to Box2D I might as well tackle some other big features and ma...| box2d.org