Neural partition of extreme-scale brain models [Theme: BrainFrame]

PDF] Cluster-Scheduling Big Graph Traversal Task for Parallel Processing in  Heterogeneous Cloud Based on DAG Transformation | Semantic Scholar

Large neural tissue models require the power of multiple computers to simulate with a practical turnaround time.  All over the model, neurons communicate with each other through a tight network of synapses.  However, inter-machine communication is slow, which limits the performance gains and scalability.  Work in this topic involves exploring algorithms that distribute large neural networks among nodes in a computer cluster to minimise inter-node communication bandwidth and delays, and measuring the actual impact of these strategies on simulation speed.

Job Category: MSc Thesis topic
Job Type: Student Thesis

Apply for this position

Allowed Type(s): .pdf, .doc, .docx