Ibero-American Symposium on Computer Graphics

Permanent URI for this community

https://diglib.eg.org/handle/10.2312/897

Browse

Browsing Ibero-American Symposium on Computer Graphics by Subject "Categories and Subject Descriptors (according to ACM CCS): I.3.2 [Computer Graphics]: Distributed/network graphics I.3.3 [Computer Graphics]: Parallel processing I.3.7 [Computer Graphics]: Raytracing"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Grid Architecture for Distributed Rendering
    (The Eurographics Association, 2006) Fernández-Sorribes, Jose Antonio; González-Morcillo, Carlos; Jiménez-Linares, Luis; Pere Brunet and Nuno Correia and Gladimir Baranoski
    Rendering is the process by means of which a raster 2D image can be obtained from the definition of a 3D scene. This process is computationally intensive and requires a lot of time to be done when the source scene has certain complexity or when high-quality realistic images are required. YAFRID (Yet Another Free Render grID) is a system that takes advantage of the characteristics of computational grids by distributing the rendering of a scene among a large number of heterogeneous computers connected to the Internet. With that kind of systems, the time a scene takes to be rendered is drastically decreased because the parts in which the complete work has been divided (called workunits) can be processed in parallel and finally joined using an interpolation function to obtain the image result. As we will discuss on the final section of this article, the selection of the workunit size is a key step in this class of work division. With this approach, local optimizations in each workunit could be made to obtain a better rendering time per frame.