An optimal parallel algorithm for volume ray casting

Volume rendering by ray casting is a computationally expensive problem. For interactive volume visualization, rendering has to be done in real time (30 frames/sec). Since the typical 3-D dataset size is 256 ³ the use of parallel processing is imperative. In this paper, we present an O(log n) EREW algorithm for volume rendering using O(n³) processors which can be optimized to O(log³ n) time using O(n³/log³ n) processors. We have implemented our algorithm on MasPar MP1200. The implementation results show that a frame from 256³ data size is generated in 11 seconds using 4096 processors. This time can be further reduced by using a large number of processors