This paper describes a scalable photorealistic renderer that is design
ed to render scenes of arbitrary complexity on computer systems of arb
itrary size. The rendering algorithm is a Monte Carlo method to comput
e approximate solutions of the rendering equation. The software implem
entation uses a diffusive load balancing method coupled with a message
driven concurrent pipeline. Measured performance in rendering replica
ted models on up to 128 computers shows scaling efficiencies as high a
s 99%. Simple extensions will partition extremely large models across
physically distributed memory as well as perform out-of-core calculati
ons.