报告题目：Energy-Efficient Computing in Mobile Platforms and Graphics Processors

报 告 人：Dr. Xin Fu, Associate Professor

主 持 人：陈铭松教授

报告时间：2018年6月12日 周二 14:00-15:00

报告地点：中北校区数学馆201

报告人简介：

Dr. Xin Fu received the Ph.D. degree in Computer Engineering from the University of Florida, Gainesville, in 2009. She was a NSF Computing Innovation Fellow with the Computer Science Department, the University of Illinois at Urbana-Champaign, Urbana, from 2009 to 2010. From 2010 to 2014, she was an Assistant Professor at the Department of Electrical Engineering and Computer Science, the University of Kansas, Lawrence. Currently, she is an Associate Professor at the Electrical and Computer Engineering Department, the University of Houston, Houston. Her research interests include computer architecture, energy-efficient computing, mobile computing, high-performance computing, deep learning, processing-in-memory, graphics processing units, hardware reliability and variability. Dr. Fu is a recipient of 2014 NSF Faculty Early CAREER Award, 2012 Kansas NSF EPSCoR First Award, and 2009 NSF Computing Innovation Fellow.

报告摘要：

This talk covers two research topics in the context of energy-efficient computing in two different platforms: i.e., mobile platforms, and graphics processors.

Mobile users always have demanding requirements on gaining an excellent user experience. There have been multiple studies on achieving the good trade-offs between QoS and energy to enhance the user experience, however, they generally ignore the fact that each individual user has his/her own preference between QoS and energy. In this talk, I will present our customized power management policy that dynamically configures the mobile platform to achieve the user-specific optimal QoS and energy trade-offs and hence, satisfy each individual mobile user.

To satisfy the ever-increasing user demands on gorgeous graphics and authentic gaming experience, today’s game developers generally employ higher image resolutions and more color effects to render 3D frames. Since 3D rendering often requires highly intensive memory access, the memory bandwidth on graphics processors becomes a severe performance and energy bottleneck. I will discuss about leveraging hybrid memory cube (HMC) technologies to enable processing-in-memory (PIM) for 3D rendering, thus, improving the energy-efficiency in graphics processors.