MISE: Providing Performance Predictability in Shared Main Memory Systems

Tuesday Feb. 19, 2013
Hamerschlag Hall D-210
4:30-5.30 pm

Lavanya Subramanian (ECE, CMU)

Abstract

Applications running concurrently on a multicore system interfere with each other at the main memory. This interference can slow down different applications differently and unpredictably. Estimating and controlling application slowdowns is necessary to achieve predictable performance, which is a critical requirement in various scenarios such as server consolidation, cloud computing, embedded systems. In this talk, I will first present a simple model to estimate application slowdowns. I will then discuss mechanisms that leverage our model to enforce guarantees on application slowdowns.

Bio

Lavanya Subramanian is a fourth year PhD student in the ECE department. She works in the broad area of computer architecture and is advised by Prof. Onur Mutlu. Her specific research interests are in providing Quality of Service in the presence of shared resources and in main memory management in multicore systems. She has a Bachelors in Electronics and Communication Engineering from Madras Institute of Technology, Anna University and a masters in ECE from Carnegie Mellon University.

Back to the seminar page