High-performance simulation cloud for development of large-scale general biophysical model

2016 Fiscal Year Final Research Report

• Project/Area Number: 26730035
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Software
• Research Institution: Osaka University
• Principal Investigator: Okita Masao 大阪大学, 情報科学研究科, 助教 (50563988)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 並列・分散処理 / 生体シミュレーション / 自動並列化

Outline of Final Research Achievements

We improved the efficiency of Flint and Flint K3, a parallel general biophysical simulator and a cloud service powered by Flint respectively, and accomplished a practical simulation of large-scale biophysical model.
Our contributions are three auto-parallelization refinement techniques: improving parallel efficiency based on two-layers graph structure, load balancing, and optimizing calculation order and data layout.
For a simulation of a 2D visual cortex neuron model that consists of more than 28 million ODEs, our method achieved 98.7% of parallel efficiency and 27.4 times speed-up.

Free Research Field

並列ソフトウェア開発