← ArchivePaper2018
LibEE 2: Enabling Fast Edits and Evaluation
Dual-representation graph engine that separates authoring nodes from evaluation tasks, achieving a 100x authoring speedup over LibEE v1 in the Premo system.
Abstract
The Premo animation platform [Gong et al. 2014] developed by DreamWorks utilized LibEE v1 [Watt et al. 2012] for high performance graph evaluation. The animator experience required fast evaluation, but did not require fast editing of the graph. LibEE v1, therefore, was never designed to support efficient edits. This talk presents an overview of how we developed LibEE v2 to enable fast editing of character rigs while still maintaining or improving upon the speed of evaluation. Overall, LibEE v2 achieves a 100x speedup of authoring operations compared with LibEE v1.
How to read this
- Category
- Production talk: a rebuilt dependency-graph engine for character rigs
- Contributions
- Presents LibEE v2, a dual-representation graph engine that separates authoring nodes from evaluation tasks
- Enables fast editing of character rigs while maintaining or improving evaluation speed
- Reports a 100x authoring-operation speedup over LibEE v1 within the Premo platform
- Context
- Direct successor to LibEE: A Multithreaded Dependency Graph for Character Animation (Watt et al. 2012), used by DreamWorks' Premo platform, where v1 prioritized evaluation speed over edit speed.Builds on: LibEE: A Multithreaded Dependency Graph for Character Animation
- Correctness
- Studio practice, not peer-reviewed; the 100x figure is a production-proven authoring speedup specific to Premo and the LibEE codebase, so it should be read as an internal engineering result rather than a generalized benchmark.
- Clarity
- Accessible to engine and pipeline readers; a single pass conveys the authoring-versus-evaluation split that motivates the redesign.
- How to read it
- Read once for the dual-representation design and why decoupling authoring from evaluation unlocks fast edits; valuable if you build or maintain a rig-evaluation graph, no formal second pass needed.
Built upon by
Related work
- LibEE: A Multithreaded Dependency Graph for Character Animation 2012 / DigiPro
- Stable and Efficient Differential IK 2010 / SIGGRAPH Asia
- Using Deep Learning to Approximate Joint Placement in 3D Bipedal Characters 2020 / Thesis
- A.C.M.E. Multilimb System 2021 / DigiPro
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →