Computing is bottlenecked by data. Large amounts of application data overwhelm storage capability, communication capability, and computation capability of the modern machines we design today. As a result, many key applications' performance, efficiency and scalability are bottlenecked by data movement. The speaker will describe three major shortcomings of modern architectures in terms of 1) dealing with data, 2) taking advantage of the vast amounts of data, and 3) exploiting different semantic properties of application data. The speaker and his research group argue that an intelligent architecture should be designed to handle data well. They show that handling data well requires designing architectures based on three key principles: 1) data-centric, 2) data-driven, 3) data-aware. The speaker will give several examples for how to exploit each of these principles to design a much more efficient and high performance computing system. He will especially discuss recent research that aims to fundamentally reduce memory latency and energy, and practically enable computation close to data, with at least two promising novel directions: 1) performing massively-parallel bulk operations in memory by exploiting the analog operational properties of memory, with low-cost changes, 2) exploiting the logic layer in 3D-stacked memory technology in various ways to accelerate important data-intensive applications. He will discuss how to enable adoption of such fundamentally more intelligent architectures, which he and his research group believe are key to efficiency, performance, and sustainability. He will conclude with some guiding principles for future computing architecture and system designs.
About the speaker
Prof. Onur Mutlu obtained his dual BSc degrees in Computer Engineering and Psychology from the University of Michigan in 2000, and his MSE and PhD in Computer Engineering from the University of Texas at Austin in 2002 and 2006 respectively. He then joined the Microsoft Research to start the Computer Architecture Group which focused on the research into memory systems for multi-core processors. He also held various product and research positions at Intel Corporation, Advanced Micro Devices, VMware and Google. In 2009, he moved to Carnegie Mellon University as an Assistant Professor of Electrical and Computer Engineering and later held the Dr. William D. and Nancy W. Strecker Early Career Professorship. He joined the ETH Zürich in 2015 and is currently a Professor of Computer Science. He is also an Adjunct Professor at Carnegie Mellon University and Bilkent University.
Prof. Mutlu’s current broader research interests are in computer architecture, computing systems, hardware security, and bioinformatics. He is especially interested in interactions across domains and between applications, system software, compilers, and microarchitecture, with a major current focus on memory and storage systems, bioinformatics, and biologically-inspired computation paradigms. A variety of techniques he and his group discovered and invented over the years have influenced the industry and have been employed in commercial microprocessors and memory/storage systems, including systems designed by Apple, Intel, IBM, Samsung, Sun Microsystems. He is the Associate Editor of Association for Computing Machinery (ACM) Transactions on Architecture and Code Optimization, ACM Transactions on Storage, and Institute of Electrical and Electronics Engineers (IEEE) Transactions on Computers.
Prof. Mutlu was named an ACM Fellow in 2017, an IEEE Fellow and an Academia Europaea Member in 2018. He has received various awards for his research, teaching and educational contributions, such as the 2019 ACM SIGARCH Maurice Wilkes Award, the 2012 Intel Early Career Faculty Honor Program Award, the 2011 IEEE Computer Society Technical Committee on Computer Architecture Young Computer Architect Award, the 2010 US National Science Foundation CAREER Award and the 2008 Microsoft Gold Star Award.
For attendees’ attention
- Light refreshments will be served from 3:30 to 4:00 pm
- Seating is on a first come, first served basis.
- In view of the Campus Access Controls currently applicable on HKUST campus, participants are required to present his/her identification document for verification before entering the campus area.
- If you are NOT a holder of a valid HKUST card (i.e. a student, staff, family, resident or temporary card of HKUST), you will need to register for attending this event by sending an email to firstname.lastname@example.org with your name displayed on your identification document at least one working day beforehand.