(L) [2013/07/31] [tby dbz] [Xeon Phi benchmarks] Wayback!For those who hadn't seen it yet, some benchmarks of the [LINK http://embree.github.io/data/embree-siggraph-2013-final.pdf Xeon Phi].
About 2x bvh build performance of an 8-core Xeon(slide 37) and about 3.5x path tracing incoherent rays performance of an 8-core Xeon(slide 45).
The 'hybrid traversal' method that is mentioned in the slides looks interesting.
(L) [2013/08/12] [tby mpeterson] [Xeon Phi benchmarks] Wayback!regarding the presentation they just compare against their own implementations.
we see 2x more rays/s for coherent transport on phi compared to embree 2 and 1.1-1.25x
for incoherent rays. phi compared to intels latest servers is in the range of 2x, so that a
dual socket server is around as fast as phi for middle-complex scenes.
mp