Ray casting on shared-memory architectures: memory-hierarchy considerations in volume rendering
Michael E. Palmer, Brian Totty, Stephen Taylor · IEEE Concurrency · 1998
To improve ray casting frame rates on shared memory architectures, the authors explore memory hierarchy effects and their interaction with parallel partitioning and load balancing. With their optimizations, a 16-processor Power Challenge machine renders a 1-Gbyte data set at a frame per second-faster than previously reported for a data set this large.