| Summary: | 博士 === 國立臺灣大學 === 資訊工程學研究所 === 102 === Ray tracing-based rendering methods can produce high quality and realistic images. However, it requires lots of computing resource to finish the rendering tasks so these kinds of methods are rarely used for small studios or individuals in the past. With the maturing of cloud computing services, small studios or individuals now rent computing resource on demand at affordable prices. This motivates us to combine the rendering with cloud computing to provide an elastic rendering framework.
In this work, we present an implementation of ray tracing with out-of-core scenes support in cloud computing environments. We design the distributed computing model, the network framework and the elastic resource utilization to support the rendering in cloud computing environments.
The experiments results show that the system has linear scalability for rendering performance with fault tolerance support. This work demonstrates the possibility for small studios or individuals to render high quality and realistic images using ray tracing-based rendering techniques in cloud computing environments. With the generalizable design, it might be also possible to extend our implementation to more applications other than rendering.
|
|---|
