I/O-Conscious Volume Rendering

Abstract

Most existing volume rendering algorithms assume that data sets are memory-resident and thus ignore the performance overhead of disk I/O. While this assumption may be true for high-performance graphics machines, it does not hold for most desktop personal workstations. To minimize the end-to-end volume rendering time, this work re-examines implementation strategies of the ray casting algorithm, taking into account both computation and I/O overhead-s. Specifically, we developed a data-driven execution model for ray casting that achieves the maximum overlap between rendering computation and disk I/O. Together with other performance optimizations, on a 300-MHz Pentium-II machine, without directional shading, our implementation is able to render a 128x128 grey-scale image from a 128x128x128 data set with an average end-to-end delay of 1 second, which is very close to the memory-resident rendering time. With a little modification, this work can also be extended to do out-of-core visualization as well.