Scientific ray tracing now can include realistic shading and material properties, but tracing rays of various depths to conclusion through partitioned data is inefficient. For such data, many ray scheduling methods have demonstrated improved rendering performance. However, synchronicity and non-adaptivity inherent in prior methods hinder further performance optimizations. In this paper, we attempt to relax these constraints. Specifically, we incorporate prediction models capable of dynamically adjusting levels of speculation in ray-data queries, making ray scheduling highly adaptable to a spectrum of scene characteristics. In addition, we organize rays in a tree of speculation nodes, where speculation is coordinated pairwise within a subtree of adaptive ray groups, facilitating concurrency and parallelism. Compared to prior non-predictive methods, we achieve up to three times higher throughput for volume and geometry rendering on a distributed system, making our method fit for both interactive and offline applications.

Computational studies of biological systems have gained widespread attention as a promising alternative to regular experimentation. Within this domain, stochastic simulation algorithms are widely used for in-silico studies of biochemical reaction networks, such as gene regulatory networks. However, inherent computational complexities limit wide-spread adoption and make traditional software solutions on general-purpose computers prohibitively slow. In this paper, we present a specialized stochastic simulation processor that exploits fineand coarse-grain parallelism in Gillepie's first reaction method to achieve high performance. The processor is designed to support large-scale networks more than a million species and reactions using external DRAMs. In addition, we introduce a dedicated compiler that creates data locality for efficient memory access and data reuse. Our performance evaluation using cycle-accurate simulation shows that our approach achieves orders of magnitude higher throughput for networks with different characteristics of coupling, compared to best-in-class software algorithms on a state-of-the-art workstation.