Xu Liu and Bo Wu received the Best Paper Award at the Supercomputing '15 conference for their paper ScaAnalyzer: A Tool to Identify Memory Scalability Bottlenecks in Parallel Programs.
The paper discusses a tool developed by the authors to improve parallel scalability of programs by identifying performance bottlenecks caused by memory subsystems. ScaAnalyzer collects memory-related metrics during execution with minimal overhead. ScaAnalyzer uses the information it collects to provide detailed, high-level guidance to programmers regarding the effects of patterns of memory access on parallel performance.
Professor Liu received his PhD from Rice University in 2014 under the supervision of John Mellor-Crummey and joined the faculty at William and Mary in Fall 2014. Professor Wu received his PhD from William and Mary in August 2014 under the supervision of Xipeng Shen and is currently a member of the EECS faculty at the Colorado School of Mines.
Reference: Xu Liu and Bo Wu, ScaAnalyzer: A Tool to Identify Memory Scalability Bottlenecks in Parallel Programs, The International Conference for High Performance Computing, Networking, Storage and Analysis (SC'15), Nov 15-20, 2015, Austin, Texas, USA.