Visualization Pipeline Performance
Visualization pipelines are commonly represented as directed graphs, but analyzing and comparing their runtime performance remains challenging. In this work, we extend pipeline representations with fine-grained runtime metrics, enabling detailed post-mortem analysis and direct visual comparison of different pipeline configurations.
By integrating performance data directly into the visual structure, our approach helps identify bottlenecks, understand the impact of pipeline changes, and support more informed optimization decisions. We demonstrate the usefulness of the approach across multiple scientific visualization scenarios.