Nov 03, 2015: Jie Shen: Efficient High Performance Computing on Heterogeneous Platforms

November 03, 2015Efficient High Performance Computing on Heterogeneous Platforms
Room: HB 2AJie Shen
12:30-13:30

Heterogeneous platforms are mixes of different processing units in a compute node (e.g., CPUs GPUs) or a chip package (e.g., APUs). This type of platforms keeps gaining popularity in computer systems ranging from supercomputers to mobile devices, and therefore improving their efficiency and usability is becoming increasingly important. The platform heterogeneity and the application diversity pose significant challenges to exploiting such platforms. In this context, systematic methods to tackle these challenges are desirable. This talk will show how we decompose this complex problem (by using an application-centric approach) and what systematic solutions we have developed (the Glinda framework and its component) for enabling a large variety of data parallel applications to achieve high performance on heterogeneous platforms.

Bio:
Jie Shen is a PhD student in the Parallel and Distributed Systems (PDS) group at Delft University of Technology. Her main research interests are in parallel computing on multi-core and many-core processors (CPUs, GPUs, and heterogeneous platforms), with a focus on performance, portability, and programmability of many-core heterogeneous platforms.