Scalable Visualization
The Scalable Visualization project researches and prototypes new techniques for porting simulations from expensive, legacy-based mainframe systems to a cluster of PC computers with high-end consumer graphics cards. The SV team is developing a heterogeneous, distributed framework that can match or even surpass the graphics processing power of more expensive platforms. By clustering PCs, the SV team is also able to create a truly scalable system, allowing unlimited resources for computationally intensive agents such as the AI-based, conversational agents found in simulations like SASO-ST. This robust, distributed framework will enable an accelerated transition of ICT training research to the Army, due to the significantly reduced cost of current-generation, commodity-based hardware.The challenge for the Scalable Visualization project is to render a seamless world view in the clustered environment, despite the graphical and temporal division of that virtual world onto separate computers. One of the main issues the project addresses is the synchronization between machines and across multiple displays. The project team has developed and tested several hardware and software solutions, although continued research and development is required.
This project differs from others in the following ways:
- It must support the Army's diverse training needs.
- It must support the sophisticated requirements of ICT projects and systems.
Tags: cluster, distributed, framework, simulation
