Viewing 0.3Tb Heart Simulation Data At Your Desk

Abstract

Atrial fibrillation (AF) affects a large section of the aged population, accounting for nearly 1% of total National Health Service (NHS) expenditure. Modelling of this condition to promote further understanding is hence of extreme importance to the NHS. Such modelling takes considerable compute power, producing large datasets; for instance, an 18 million data point model produces 0.3Tb when simulating 3 seconds of heart activity. Such large datasets are currently batch processed to produce a movie from a preset camera position. What is needed is a means of reducing the data to a more manageable size, where it can be displayed in real-time at the heart modeller s desk for interactive exploration and evaluation. We present a pilot study where the Resource-Aware Visualization Environment (RAVE) has been adapted to filter AF datasets ready for subsequent interactive visualization on a desktop machine. A high-end visualization workstation was not required to view the 0.3Tb dataset in real-time, where the modeller could manipulate both camera position/orientation and simulated elapsed time at interactive rates. This has enabled visual inspection and related rapid evaluation of AF, as well as supporting intuitive insight.