Performance analysis of a virtual reality development environment: Measuring and tooling performance of VR Juggler

Abstract

Performance is a critical concern in the implementation of interactive Virtual Reality (VR) programs. Many standalone performance monitoring tools are poorly suited to analyzing complex, multithreaded, interactive VR applications. Integrating performance measurements into a VR application toolkit can provide developers with useful, complete, and understandable performance information, which can then be used to analyze and optimize their applications. This thesis covers the design and implementation of an extensible performance monitoring system for VR programs. The data-collection capabilities are integrated into the VR Juggler development system for VR applications. Also included is a tool for interactively visualizing performance data gathered from running applications. As proof of these performance tools' utility, a battery of tests are described in this document, showing how performance monitoring has been used to debug and optimize the VR Juggler libraries, and also demonstrating how performance measurements can be used by application developers to optimize their programs.