A software toolkit for dynamic visual computation on large-scale ultra-high-resolution multi-display setup with distributed GPU architecture

The state-of-the-art multi-display setups powered by distributed GPU architectures (DGA) open significant potential for visual creation and scientific visualization by providing an ultra-high resolution computational display surface.

However, as a DGA consists of many distributed computers, complex network operations are required to utilize all the screens as a unified display area. This in turn creates an accessibility barrier, since the users of such a system have to deal with multiple networked computers, an array of synchronization and rendering problems and closed and obtuse system integration solutions. These factors seriously block a whole set of creators from using DGA displays as a productive visual instrument.

Chameleon is a universal toolkit for dynamic visual computation on multi-display setups powered by distributed computers. The toolkit is designed to include only the most-necessary tools for DGA computation. With it, one can match existing content created for a single computer and display it on a multi-display system with DGA. Moreover, the toolkit provides a seamless user experience of cross-computer operations by providing a graphical user interface (GUI), where the user can group any or all screens as one rendering surface and attach desired content to it.

Users can also directly adjust the content with built-in control signals on the fly. After the user finishes the configuration of all screens, he/she can save the current composition as a preset. The user is then able to switch back and forth between different saved presets. Additionally, he/she can create a timeline of presets to enable the system to automatically loop through them.

Chameleon is the first step of development aimed at enabling the use of DGA systems by an expanded user audience.

*Chameleon is my graduate project for M.S. in Media Arts and Technology, UC Santa Barbara