FLAME GPU: Complex System Simulation Framework
Authors: Mozhgan Kabiri Chimeh (University of Sheffield)
Abstract: Agent based modeling can be used to represent complex systems at differing biological scales
FLAME GPU (Flexible Large Scale Agent Based Modelling Environment) :
• Aim: High performance agent based simulation without writing lots of GPU code
• Technique: Agents are state machines with memory (communicating X-Machines)
o Uses state based agents to automatically generate CUDA GPU simulation code
• Optimisations: Reduce divergence and use efficient data structures/algorithms for common agent based behaviours
• Scope: Large numbers of relatively simple discrete time step agents
• Examples:
o Swarm simulations
o Cellular level simulation of epithelium tissue growth
o Pedestrian simulation
Archive Materials
Back to Women in HPC: Diversifying the HPC Community Archive Listing
Back to Full Workshop Archive Listing