Tutorial

HPCS 2017 FLAME GPU Tutorial - 16th July 2017

Welcome to the FLAMEGPU Tutorial!

This tutorial will describe parallelisation techniques for problems expressed as complex system using a generalised multi-agent approach. Attendees will gain direct hands on experience of targeting advanced GPU architectures in the cloud by using the open source FLAME GPU framework.

Tutorial Format
Half a day (4 hours).

Tutorial Outline
We cover the following topics: 
  • Implications of highly parallel architectures to multi-agent systems (20 min) 
  • Parallelising multi-agent models:  The FLAME GPU methodology (40 min)
  • Hands-on experience with FLAME GPU framework (120 min excluding break)
  • Discussion, questions and feedback (30 min)

Link to Slides

OTHER LINKS:

Link to Github (starting code and solutions)
Link to tutorial proposal document
Comments