Tyson Ibele on Tyflow
In the latest Post-Post Podcast, David Gidali interviews Tyson Ibele about his life, career, games and, of particular interest to 3DS Max users, TyFlow – his eagerly anticipated new simulation software. Listen on the Post-Post website. TyFlow discussion starts in earnest at about the 1 hour point
Tyson has also released a couple more demos in action recently. In this first one we see how the cloth solver handles large amounts of input geometry.
tyFlow's OpenCL-accelerated cloth solver can handle huge amounts of input geometry, and since each vertex on a converted cloth mesh is maintained as a regular tyFlow particle, those individual vertices can be manipulated like any other particles within a flow. In this example, 350 pieces of clothing with over 7 million combined faces are controlled with wind, gravity and surface attraction forces. In total, 65 million mass-spring bindings are required to simulate all of the cloth physics and inter-particle collision forces. With 75 per-step cloth solver iterations and 4 overall solver steps per frame, that's 450 billion constraint evaluations required for every second of animation you see here. Despite the huge number of computations required to solve the system, the simulation was cached on a multi-core machine with a GeForce 1080ti GPU in only 1.5 minutes per frame. That kind of power and speed makes tyFlow an obvious choice for high resolution offline simulations. #tyflow #autodesk #physx #procedural #generative #animation #cloth #rigidbody #softbody #simulation #cg #3d #vfx #3dsmax #particles #mdcommunity #mgcollective #ssequential #plsur #chaosgroup #vray
And in the second we see a demo that combines different forces and effects.