Energy-efficient scheduling for heterogeneous parallel computing environments

在本研究中，为了降低高性能计算的能耗，我们的目标是设计实现高吞吐量计算和降低能耗的方法。首先，为了实现高性能计算，我们开发了构建大规模计算网格的方法。接下来，我们把大规模计算网格看作是大规模分布式系统，并提出了在大规模分布式系统中降低能耗的方法。此外，我们专注于移动的代理来管理大规模的计算网格，并开发方法来控制移动的代理与低通信成本。

In this research, to reduce energy consumption in high-performance computing, we aim to design methods that attain high-throughput computing and reduce energy consumption. First, to achieve high-performance computing, we develop methods to construct large-scale computing grids. Next, we regard large-scale computing grids as large-scale distributed systems, and develop methods to reduce energy consumption in large-scale distributed systems. In addition, we focus on mobile agents to manage large-scale computing grids, and develop methods to control mobile agents with low communication costs.