ExaViz: a flexible framework to analyse, steer and interact with molecular dynamics simulations.

Area of Science:

• Computational biology
• Molecular dynamics simulations
• Scientific visualization

Background:

• Large-scale molecular dynamics simulations generate massive datasets, overwhelming traditional offline analysis methods and supercomputer I/O capabilities.
• Current post-processing and analysis workflows are cumbersome, requiring significant data storage and remote handling from scientists' offices.

Purpose of the Study:

• To introduce the ExaViz framework, an alternative approach for analyzing, steering, and interacting with molecular dynamics simulations.
• To enable in-situ data processing, reducing the volume of saved data and alleviating I/O bottlenecks.

Main Methods:

• Coupling molecular dynamics simulations with analysis tools for real-time data processing at the source.
• Developing interactive and batch session capabilities for diverse simulation lengths.
• Integrating visualization tools, haptic devices, and immersive CAVE-like environments for user interaction.

Main Results:

• ExaViz processes data close to its creation, generating smaller, pre-processed datasets.
• The framework supports both live and batch simulation sessions with real-time analysis and steering.
• Interactive sessions allow users to visualize molecular systems, display analysis plots, and control simulations using haptic devices.

Conclusions:

• ExaViz offers an efficient alternative to traditional offline analysis for large-scale molecular dynamics simulations.
• The framework enhances user interaction and data management, improving the analysis workflow.
• Immersive environments can further leverage ExaViz for intuitive navigation and visualization of complex molecular data.