新闻动态

    2013年7月2日，美国罗切斯特大学Ding Chen教授来访中心并作了题为“Modern Locality Theories and Applications”的学术报告。

Abstract: In computing, there is a fundamental conflict between computing speed and data capacity. A signal cannot travel faster than the speed of light, so the faster the data access is, the smaller amount of data it can reach. The solution is to trade space for speed and store data in a memory hierarchy. Locality theory is concerned with the fundamental properties of data usage in applications and memory management in hardware, VM, OS, and other run-time systems.

I'll review multiple branches of the past research in different areas that culminate in our higher-order theory [ASPLOS'13] showing that the locality metrics for performance evaluation, system management, and
program optimization --- hitherto separately developed and applied --- are eseentially equivalent. Under this new light, I'll discuss the current advances in multicore/GPU cache management, collaborative caching, cache conscious algorithms, and some recent interests related to BigData.

Biography: Ding, Chen graduated from Beijing University in 1994 and is currently Professor of Computer Science at University of Rochester. His research received the young investigator awards from NSF and DOE. He co-founded the ACM SIGPLAN Workshop on Memory System Performance and Correctness (MSPC) and was a visiting researcher at Microsoft Research and a visiting associate professor at MIT. He is currently a visitor in Professors Luo, Yingwei and Wang, Xiaolin's group at Beijing University.