Lier, AlexanderStamminger, MarcSelgrad, KaiPatney, Anjul and Niessner, Matthias2018-11-112018-11-112018978-1-4503-5896-52079-8679https://doi.org/10.1145/3231578.3231583https://diglib.eg.org:443/handle/10.1145/3231578-3231583In this paper we describe and evaluate an implementation of CPUstyle SIMD ray traversal on the GPU. We show how spreading moderately wide BVHs (up to a branching factor of eight) across multiple threads in a warp can improve performance while not requiring expensive pre-processing. e presented ray-traversal method exhibits improved traversal performance especially for increasingly incoherent rays.Computing methodologiesRay tracingShared memory algorithmsVector / streaming algorithmsRay tracingGPUSIMD10.1145/3231578.3231583