This workshop will focus on Multi GPU System Design, implementation and execution with a focus on stability, performance and simplicity. Participants will examine the inner workings of a custom multi-node playlist component (Fusion Player)that plays back a playlists of cues (either pre-rendered content files or real-time components) across multiple nodes (ie. GPU's and or servers). In the workshop, methods for communication, synchronization and configuration of each node in the network will be discussed along with various methods and techniques to keep performance high while at the same time maintaining high functionality along with high render quality.
TD-centric advanced Python and GLSL techniques used to maintain simplicity and speed will also be covered.
At the end of the workshop attendees should have a solid understanding of the requirements of implementing and executing a large scale system as well as an understanding of the data model flow of the Fusion Player component itself. They will also be able to keep the Fusion Player component for future use and reference.