Ray tracing on distributed memory parallel computers : strategies for distributing computations and data

The ray tracing algorithm produces high quality images, but it requires a lot of computations which makes it extremely time-consuming. Several attempts have been made to reduce the synthesis time by using high speed parallel computers. For example, distributed memory parallel computers, such as hypercubes or transputer-based machines, offer an interesting performance/cost ratio. Several studies have used parallel computers to speed up the ray tracing algorithm so as to render complex scenes with several thousands of objects. However few of them have led to real implementation due to the complexity of the solution proposed by the authors. This paper introduces a new approach emulating a read-only shared memory on a distributed memory parallel computer. Results are given and are compared to a previous parallel ray tracing algorithm, both are implemented on an iPSC/2