A technique for workload balancing on distributed visualization clusters is suggested. The round-robin scheduling of partial object geometry "chunks" avoids render node starvation and balances the data backl...
详细信息
ISBN:
(纸本)0889863377
A technique for workload balancing on distributed visualization clusters is suggested. The round-robin scheduling of partial object geometry "chunks" avoids render node starvation and balances the data backlog in the send buffers of the geometry nodes. Because frame synchronization requires that all send buffers drain, balancing these send buffer backlogs is seen to substantially improve frame rate performance for dynamic displays.
We describe a novel distributed graphics system that allows an application to render to a large tiled display. Our system, called WireGL, uses a cluster of off-the-shelf PCs connected with a high-speed network. WireGL...
详细信息
ISBN:
(纸本)9780780398023
We describe a novel distributed graphics system that allows an application to render to a large tiled display. Our system, called WireGL, uses a cluster of off-the-shelf PCs connected with a high-speed network. WireGL allows an unmodified existing application to achieve scalable output resolution on such a display. This paper presents an efficient sorting algorithm which minimizes the network traffic for a scalable display. We will demonstrate that for most applications, our system provides scalableoutput resolution with minimal performance impact.
A stable frame rate is important for interactive rendering systems. Image-based modeling and rendering (IBMR) techniques, which model parts of the scene with image sprites, are a promising technique for interactive sy...
详细信息
ISBN:
(纸本)0819444057
A stable frame rate is important for interactive rendering systems. Image-based modeling and rendering (IBMR) techniques, which model parts of the scene with image sprites, are a promising technique for interactive systems because they allow the sprite to be manipulated instead of the underlying scene geometry. However, with IBMR techniques a frequent problem is an unstable frame rate, because generating an image sprite (with 3D rendering) is time-consuming relative to manipulating the sprite (with 2D image resampling). This paper describes one solution to this problem, by distributing an IBMR technique into a collection of cooperating threads and executable programs across two computers. The particular IBMR technique distributed here is the LOD-Sprite algorithm. This technique uses a multiple level-of-detail (LOD) scene representation. It first renders a keyframe from a high-LOD representation, and then caches the frame as an image sprite. It renders subsequent spriteframes by texture-mapping the cached image sprite into a lower-LOD representation. We describe a distributed architecture and implementation of LOD-Sprite, in the context of terrain rendering, which takes advantage of graphics hardware. We present timing results which indicate we have achieved a stable frame rate. In addition to LOD-Sprite, our distribution method holds promise for other IBMR techniques.
The paper describes the design, implementation, and use of a commercial metacomputing environment for computationally intensive loosely-coupled parallel applications. Much weight has been laid on practical and commerc...
详细信息
The paper describes the design, implementation, and use of a commercial metacomputing environment for computationally intensive loosely-coupled parallel applications. Much weight has been laid on practical and commercialisation aspects, and on business benefit. This distinguishes this work from many other metacomputing activities in a positive way. It demonstrates how a metacomputing environment can be used to improve a company's position in the market. A cluster of networked geographically dispersed computing nodes is considered as physical layer. The proposed distribution of work over the nodes of the execution network is proven optimal, in terms of minimizing the execution time, with respect to the availability of resources. We also present our experience on testing the environment for computing-intensive 3D-rendering jobs derived from the ESPRIT project EROPPA and demonstrate that the new environment can change dramatically the character of the post production business. (C) 1999 Published by Elsevier Science B.V. All rights reserved.
暂无评论