DataSys: Data-Intensive Distributed Systems LaboratoryData-Intensive Distributed Systems Laboratory

Illinois Institute of Technology
Department of Computer Science

Towards Many-Task Computing on Exascale 

Zhao Zhang

PhD Student
Computer Science Department
University of Chicago

In this talk, we will discuss the issues that Many-Task Computing on the next generation of Supercomputers, such as Mira, the IBM BG/Q deployment. The topics range from resource provisioning, task management, collective data management, file system up to programming language and application design of MTC. I will share some of the real experience working on large systems with both conceptual and tech issues. The work is primarily based on my master thesis, AME, an any-scale Many-Task Computing Engine.