Alexandra "Sasha" Fedorova

My data or yours? Orchestrating the movement and placement of data on large multicore systems 02-26-2015 @ 2:30 - 3:30

Alexandra Fedorova — Alexandra (Sasha) Fedorova is an Associate Professor of Computer Science at Simon Fraser University. She earned her Ph.D. at Harvard in 2006. Her thesis was about designing new OS scheduling algorithms for multicore systems that improve the management of crucial shared resources, such as caches and memory bandwidth. While completing PhD she interned at Sun Labs where she co-authored the simulator for the Sun's multicore processor Niagara. In 2006 Sasha joined the School of Computing Science at SFU where she co-founded the Systems, Networking and Architecture (SYNAR) research lab and currently leads a group of 12 graduate students and postdocs. Sasha has more than 30 publications in top scientific venues and her research is supported by the NSERC, British Columbia Innovation Council, Oracle, Google, Intel, ST Microelectronics, Research in Motion and Electronic Arts. Sasha is the recipient of the 2011 Anita Borg Early Career Award, and in 2012 she was named an Alfred P. Sloan Fellow.

Data movement is becoming a major bottleneck in all kinds of programs running on multicore systems. A program accessing data over a congested cross-chip interconnect could run as much as 3 times slower than a program running without congestion. Adding a shared counter update to a scalable parallel key-value store could slow down response time by as much as 4x. And yet, data movement is inevitable, simply because programs become more parallel and access more data.

In my research group we are designing a range of techniques to mitigate the bottlenecks associated with data movement: from operating system page-placement algorithms, new programming libraries and language extensions, to tools that analyze and visualize data access patterns. In this talk, I will highlight our contributions in this space.


Video (46:31)