OpenSubdiv is a set of open source libraries that implement high performance subdivision surface (subdiv) evaluation on massively parallel CPU and GPU architectures. This codepath is optimized for drawing deforming subdivs with static topology at interactive framerates. The resulting limit surface matches Pixar’s Renderman to numerical precision.
OpenSubdiv is covered by the Microsoft Public License, and is free to use for commercial or non-commercial use. This is the same code that Pixar uses internally for animated film production. Our intent is to encourage high performance accurate subdiv drawing by giving away the “good stuff”.
Note that this beta code is live and will undergo significant churn as it approaches release. Expect that APIs will change as the code is continually improved.
There are many things we'd love to do to improve support for subdivs but don't have the resources to. We hope folks feel welcome to contribute if they have the interest and time. Some things that could be improved:
* The reference maya plugin doesn't integrate with Maya shading. That would be cool.
* John Lasseter loves looking at film assets in progress on an iPad. If anyone were to get this working on iOS he'd be looking at your code, and the apple geeks in all of us would smile.
* Alembic support would be wonderful, but we don't use Alembic enough internally to do the work.
* The precomputation step with hbr can be slow. Does anyone have thoughts on higher performance with topology rich data structures needed for feature adaptive subdivision? Maybe a class that packs adjacency into blocks of indices efficiently, or supports multithreading, or even feature-adaptive subdivision on the GPU?