Yet Faster Ray-Triangle Intersection (Using SSE4)

Ray-triangle intersection is an important algorithm, not only in thefield of realistic rendering (based on ray tracing), but also inphysics simulation, collision detection, modelling, etc. Obviously, thespeed of this well-defined algorithm's implementations is importantbecause calls to such a routine are numerous in rendering andsimulation applications. Contemporary fast intersection algorithms,which use SIMD instructions, focus on the intersection of ray packetsagainst triangles. For intersection between single rays and triangles,operations such as horizontal addition or dot product are required. TheSSE4 instruction set adds the dot product instruction which can be usedfor this purpose. This article presents a new modification of the fastray-triangle intersection algorithms commonly used, which - whenimplemented on SSE4 - outperforms the current state-of-the-artalgorithms. It also allows both a single ray and ray packetintersection calculation with the same precomputed data. The speed gainmeasurements are described and discussed in the article.