Thomas Wenisch selected as Maurice Wilkes Award Recipient

The award recognizes Prof. Wenisch’s contributions to memory persistency and energy-efficient systems.

Thomas Wenisch Enlarge
Prof. Thomas Wenisch

Prof. Thomas F. Wenisch has received the ACM-SIGARCH Maurice Wilkes Award, the most prestigious mid-career award in the computer architecture community, for his “contributions to memory persistency and energy-efficient systems.”

The award is named for Sir Maurice Wilkes, the Director of the University of Cambridge Computer Laboratory during the development of stored program computers, and inventor of many computing concepts still in use today. It is presented at the International Symposium on Computer Architecture.

Wenisch’s research has investigated multiprocessor and multicore systems, multicore programmability, data center architecture, and performance evaluation methodology. His work seeks to enhance the performance and reduce the energy consumption, capital costs, and carbon footprint of cloud infrastructure by improving efficiency at all scales, from client devices and server systems to facility-scale power and cooling infrastructure.

Wenisch’s interests in energy efficient computing, data centers, and warehouse scale computing date back to 1998, well before he entered graduate school. He held an undergraduate internship at American Power Conversion in West Kingston, Rhode Island, where he was first exposed to studies on data center thermal topology estimation. During his time in graduate school, Wenisch saw the beginning of the transformation from grossly inefficient data centers to the hyperscale infrastructure of today’s cloud computing.

Wenisch focused his early career on energy efficient enterprise systems and making servers “green.” In their first paper together, he and his first PhD student, David Meisner, introduced the idea of PowerNap, a plan to essentially put servers to sleep for milliseconds between tasks. This kindled interest in a variety of follow-ups to develop techniques to make servers sleep more effectively when idling. After Meisner completed an internship at Google, where he was able to observe server power management techniques at Google scale, Wenisch and Meisner shifted the focus of their work from idle low power modes to active low power modes as increasing core counts and multithreading made it harder to find opportunities to put a large fraction of a system to sleep. The work on active low power modes grew into a multi-year set of collaborations that focused on main memory, the next largest memory hog after the CPU.

Wenisch also collaborated with Milo Martin at the University of Pennsylvania on a notable project called Computational Sprinting. While conventional processor designs (including their energy delivery systems and heat sinks) are designed primarily for sustained performance, the researchers asked the question: “What would a system look like if designed to provide responsiveness during bursts rather than with a singular focus on sustained performance?” Their approach was aimed at mobile devices, where many interactive applications are characterized by short bursts of computational demand punctuated by long idle periods waiting for user input. Computational sprinting activates otherwise powered-down cores for sub-second bursts of intense parallel computation in response to such sporadic user activity. During sprints, the processor generates heat at a rate that far exceeds the thermal (cooling) and electrical (power delivery and stability) capacities of a typical mobile device, but for just a short period of time. Phase-changing materials were used to absorb the heat from these sprints.

The selection committee also highlights Wenisch’s work on “memory persistency” in its award citation. This line of work examines how to design software systems that can best take advantage of new memory technologies that save their content when power is lost (like disks), but can also be accessed directly by processor instructions (like memory). Memory persistency lays out principles on how memory accesses to persistent memory devices must be ordered to build software systems that can recover after power failures or crashes.

Wenisch has been the principal investigator/theme leader for the Algorithm-Driven Architectures theme within the Applications Driving Architectures (ADA) Research Center, a $32 million center based at Michigan which aims to streamline and democratize the design and manufacturing of next-generation computing systems. He has held this position since 2018, when the center launched. The ADA Center seeks to design innovative cross-layer approaches to accelerate important applications and algorithms in an era where Moore’s law and Silicon scaling alone won’t continue to improve computer system performance.

Wenisch received his PhD in electrical and computer engineering from Carnegie Mellon University in 2007 and joined the faculty at Michigan that year. He was the recipient of an NSF CAREER award in 2009 and was named a Morris Wellman Faculty Development Assistant Professor of EECS in 2011. He received both the U-M Henry Russel Award for academic accomplishment and the College of Engineering Ruth and Joel Spira Teaching Award in 2013. He is listed in the MICRO Hall of Fame, the ISCA Hall of Fame, and the HPCA Hall of Fame for publishing at least eight papers at each of those conferences.