Interactive High-Resolution Isosurface Ray Casting on Multicore Processors

We present a new method for the interactive rendering of isosurfaces using ray casting on multicore processors. This method consists of a combination of an object-order traversal that coarsely identifies possible candidate three-dimensional (3D) data blocks for each small set of contiguous pixels and an isosurface ray casting strategy tailored for the resulting limited-size lists of candidate 3D data blocks. Our implementation scheme results in a compact indexing structure and makes careful use of multithreading and memory management environments commonly present in multicore processors. Although static screen partitioning is widely used in the literature, our scheme starts with an image partitioning for the initial stage and then performs dynamic allocation of groups of ray casting tasks among the different threads to ensure almost equal loads among the different cores while maintaining spatial locality. We also pay a particular attention to the overhead incurred by moving the data across the different levels of the memory hierarchy. We test our system on a two-processor Clovertown platform, each consisting of a Quad-Core 1.86-GHz Intel Xeon Processor and present detailed experimental results for a number of widely different benchmarks. We show that our system is efficient and scalable and achieves high cache performance and excellent load balancing, resulting in an overall performance that is superior to any of the previous algorithms. In fact, we achieve interactive isosurface rendering on a screen with 1.0242 resolution for all the data sets tested up to the maximum size that can fit in the main memory of our platform.