Abstract
Steady and rapid improvement of computing technology enables a simulation with very high resolution. Demands to visualize a huge amount of volume data with internal structure and dynamic behavior increase in spite of high rendering costs and large memory occupancy. On the other hand, commodity clusters based on consumer PCs are familiar to scientists and engineers of computational sciences. This paper describes PC-based implementation of volume visualization processes for large-scale simulation data sets, aiming at the development of commodity-cluster-based visualization system. The visualization flow consists of three stages; remapping stage, transfer function stage, rendering stage. Implementation and computational performance of each stage are discussed in detail.
