[2020-01-03] Prof. Hung-Wei Tseng, University of California, Riverside, "Programming in the dark silicon era"

Poster:Post date:2019-12-05
Title: Programming in the dark silicon era
Date: 2020-01-03 3:40pm-5:00pm
Location: R102, CSIE
Speaker: Prof. Hung-Wei Tseng, University of California, Riverside
Hosted by: Prof. Hsin-Mu (Michael) Tsai


We’re now in the dark silicon era that the performance of general-purpose processors, no matter single-core or multicore ones, do not get any faster. Dark silicon has significantly changed the landscape of computing as we see the emergence of domain-specific, heterogeneous computing (e.g., GPUs, TPUs) devices to deliver the desired computation throughput. However, unlike conventional micro-processors that programmers can treat as black boxes most of the time, programmers can easily abuse these emerging devices and lead to unwanted performance or even incorrect results.
Comparing with micro-processors, heterogeneous accelerators are usually less precise, throughput-oriented instead of latency sensitive. In this talk, Hung-Wei will give examples on how the characteristics would lead to suboptimal performance or make the resulting program incorrect. Hung-Wei will also discuss their latest publication on how to address the potential performance issues in systems with these accelerators.
To further release the burden of the programmer in heterogeneous computers, Hung-Wei will also introduce their latest, on-going research project -- a dynamic language framework for heterogeneous computers with near-data accelerators. By taking advantage of dynamic languages’ runtime code generation/optimization as well as the backend just-in-time compilation, the resulting framework can speedup applications by up to 1.45x.
Hung-Wei is currently an assistant professor in the Department of Electrical and Computer Engineering at the University of California, Riverside. He is now leading the Extreme Storage & Computer Architecture Laboratory and focusing on tackling the performance issues in modern heterogeneous computer systems through intelligent data storage. He is recognized by facebook faculty research award for his research in accelerating data-intensive applications through revisiting the storage system design. Prior to joining UCR, Hung-Wei was an assistant professor at NC State University and a postdoctoral scholar in the Department of Computer Science and Engineering at the University of California, San Diego. He got his PhD from Department of Computer Science and Engineering at the University of California, San Diego. His thesis work with Professor Dean Tullsen, data-triggered threads, was selected by IEEE Micro "Top Picks from Computer Architecture" in 2012.
Last modification time:2019-12-05 AM 10:06

cron web_use_log