CS set to honor four distinguished faculty with named chairs and professorships

11/14/2016 Laura Schmitt, CS @ ILLINOIS

One of the highest honors the campus can bestow, named chairs and professorships acknowledge outstanding faculty research, service, and education accomplishments.

Written by Laura Schmitt, CS @ ILLINOIS

One of the highest honors the campus can bestow, named chairs and professorships acknowledge outstanding faculty research, service, and education accomplishments. In the coming weeks, four CS faculty—Sarita V. Adve, David A. Forsyth, Laxmikant “Sanjay” Kale, and Josep Torrellas—will receive these titles thanks to the generosity of CS alumni and friends.

On November 29, Sarita Adve will be invested as the Richard T. Cheng Professor in Computer Science, while David A. Forsyth will be invested as the Fulton Watson Copp Chair in Computer Science. On December 15, the department will celebrate the investiture of Laxmikant “Sanjay” Kale as the Paul and Cynthia Saylor Professor in Computer Science and the investiture of Josep Torrellas as the Saburo Muraga Professor in Computer Science.

Sarita V. Adve
Sarita V. Adve

About the Richard T. Cheng Professorship

An influential educator who founded the computer science departments at the Rochester Institute of Technology and Old Dominion University, Richard T. Cheng (MS CS '69, PhD '71) established a professorship in 2015 to help CS @ ILLINOIS retain and attract talented faculty. Earlier in his career, Cheng founded ECI Systems & Engineering. At its peak, the Virginia-based company was a leading integrated systems provider for the U.S. military and government, with annual sales of $50 million, more than 500 employees, and 32 offices around the world. In 2011, Cheng received the College of Engineering at Illinois Alumni Award for Distinguished Service for his outstanding leadership in education and in business.

Sarita V. Adve: The Richard T. Cheng Professorship in Computer Science

For years, the memory consistency model, which affects a computer’s programmability and performance, was one of the most challenging and contentious areas in concurrent hardware and software specification. One solution, sequential consistency, was the simplest to program, but most systems did not provide it for performance reasons.

Instead, the solution was to have divergent models—often ambiguously specified—for different hardware. That is, until CS Professor Sarita V. Adve helped bring the hardware and software communities together to address the problem.

Adve’s early work, with her advisor Mark Hill, departed from the prevalent hardware-centric approaches to use a combined hardware/software view more appropriate for an interface. She observed that for well-synchronized programs, which she formalized as data-race-free, both sequential consistency and high performance could be achieved. That work observed that for well-synchronized programs, formalized as data-race-free, both sequential consistency and high performance could be achieved. Over several years, Adve worked closely with hardware and software researchers and practitioners to forge consensus towards adopting the data-race-free model as the standard. Today, this model is the foundation of the memory models for most of the popular programming languages such as Java, C++, and C.

In collaboration with industry and other Illinois researchers, Adve has also explored hardware resiliency.  Her SWAT project was one of the first to develop a very low-cost comprehensive resiliency framework that detects, diagnoses, and recovers from several fault types. This work is often credited for making software-driven solutions widely accepted as a promising approach for hardware resiliency.

Adve has also made key contributions to low-power systems. She and her collaborators at Illinois were among the first to recognize that significant power reductions required breaking traditional system boundaries in favor of a cross-layer system-wide power management framework. Her GRACE project was the first to demonstrate a prototype where the hardware, operating system, network, and applications adapted collaboratively to minimize energy while meeting quality-of-service requirements.

Adve’s recent work, with Vikram Adve, has challenged the research community to rethink the design of parallel languages and hardware. Inspired by trends in disciplined parallel programming, her DeNovo project has developed cache coherence protocols and communication architectures that are simpler, higher performance, and lower energy.

A member of the CRA Board of Directors and chair of ACM SIGARCH, Adve is a Fellow of both ACM and IEEE. She is also a recipient of the Anita Borg Institute Women of Vision Award in the Innovation category and the ACM SIGARCH Maurice Wilkes Award.

David A. Forsyth
David A. Forsyth

About the Fulton Watson Copp Chair

Established with a gift from the namesake’s estate, the Fulton Watson Copp Chair recognizes a faculty member who is an internationally renowned leader in computer science, has an ongoing research program central to the mission of the department, and is a prominent educator with a reputation for outstanding, innovative teaching.

Fulton Copp earned his BS in electrical engineering in 1925. He was a member of the ROTC and served in the U.S. Calvary. Early in his career he managed a gold mine in the Sierras. During World War II, he served in the Corps of Engineers and was involved in the construction of Army hospitals and airfields in this country and the Pacific. He retired as a Lt. Col. from the reserves in 1962.

After the war, Mr. Copp was involved in the construction and management of natural gas pipelines in the U.S., Australia, and Canada. He was also worked on oil extraction from shale. He retired in 1972 from the Bechtel Corporation. Mr. Copp died in 1990.

David A. Forsyth: The Fulton Watson Copp Chair in Computer Science

A leading researcher in computer vision, CS Professor David A. Forsyth has made distinctive contributions to human motion computing (detecting, understanding, and animating what people do), to how computers relate words and pictures, and to rendering objects into photographs.

Forsyth's group started the trend of attaching words to images by developing an award-winning model of object recognition as machine translation that could annotate image regions with words. More recent work in this area offered the first method to produce sentences that describe images; this, too, is now a hot topic. In collaboration with Professor Derek Hoiem, also of Illinois, Forsyth's group showed how to describe unfamiliar objects in pictures by computing their attributes. This is now a standard strategy in object recognition.

Forsyth's group has studied methods to analyze pictures and movies of people. He wrote important early papers which showed how to tell whether there were human nudes in a picture, now an important practical application. The group developed the first robust, accurate human tracker that can reliably report the configurations of arms and legs, and which does not need to be started by hand. The group also developed widely-cited methods for rearranging motion-capture data to produce highly-realistic human animations.

Forsyth’s recent work on realistically rendering synthetic objects into legacy photographs was widely covered (e.g., Wired, The Atlantic, Popular Science, New Scientist, and IEEE Spectrum), and has resulted in three patents.

Forsyth is a highly visible educator in computer vision. His book, “Computer Vision: A Modern Approach,” (with Jean Ponce), is widely adopted as a course text, since it provides a unified vision of the field. Many ex-students are influential professors in computer vision, and two of his ex students have been awarded the Marr prize for work as professors.

Currently in a second term as Editor-in-Chief for IEEE Transactions on Pattern Analysis and Machine Intelligence, Forsyth has regularly served as a program or general chair for the top conferences in computer vision.

A Fellow of the ACM (2014) and IEEE (2009), Forsyth has also been recognized with the IEEE Computer Society’s Technical Achievement Award (2005), the Marr Prize, and a prize for best paper in cognitive computer vision (ECCV 2002).

Laxmikant "Sanjay" Kale
Laxmikant "Sanjay" Kale

About the Paul and Cynthia Saylor Professorship

The Paul and Cynthia Saylor Professorship in Computer Science was established in 2005 through the generosity of the late Gene H. Golub (B.S. Mathematics, 1953, M.S. Mathematical Statistics, 1954, Ph.D. Mathematics, 1959) in honor of his longstanding friendship with the Saylors.

Golub's influence on the theory and practice of scientific computing was profound, both through his vast range of technical research contributions and his professional leadership of the entire scientific computing community nationally and internationally.

Golub was a cofounder of the Stanford computer science department, and served on the Stanford faculty beginning in 1962 through the time of his death in November of 2007. He was a member of both the National Academy of Science and the National Academy of Engineering. His founding of NA-Net and NA-Digest helped unify the worldwide numerical analysis community.

Laxmikant “Sanjay” Kale: The Paul and Cynthia Saylor Professorship in Computer Science

An expert in parallel programming, CS Professor Laxmikant “Sanjay” Kalé has developed tools and abstractions that make parallel computing easier and more efficient for scientists and engineers who model complex problems in epidemiology, quantum chemistry, biophysics, and astronomy.

He pioneered the idea of a powerful, introspective and adaptive runtime system to simplify parallel programming of complex applications and to automate resource management. To enable such adaptivity, he developed a class of parallel programming models based on over-decomposition, migratability, and asynchrony. His research group developed the Charm++ parallel programming system embodying these ideas. Charm++ has a 20- year track record of delivering high performance on systems ranging from single workstations to the largest supercomputers in the world. Charm++ is one of the few academically developed parallel programming systems that is used to solve real-world problems.

Using Charm++, he has developed many collaborative parallel applications, including Nanoscale Molecular Dynamics (NAMD), with the late Illinois Biophysics Professor Klaus Schulten. NAMD was used for the world’s first simulation of the precise chemical structure of the HIV capsid—a major breakthrough in the understanding of the deadly virus.

Kalé’s group has collaborated to develop applications such as ChaNGa, which enables astronomers to study the origins and evolution of the universe, and OpenaAtom, which enables studies of quantum-mechanical details of photovoltaic materials, which may lead to better solar cells. The EpiSimdemics project simulates the spread of contagions like the H1N1 and Ebola viruses as they propagate through populations and the impact of potential interventions on the spread of the contagion.

Prof. Kalé’s research interests include higher level parallel programming abstractions, dynamic load balancing, fault tolerance, energy optimization in parallel computing, structured and unstructured mesh computations with dynamic refinements, parallel discrete event simulation, simulations of future parallel machines, performance analysis, and sustainable development of software in academia.

A fellow of IEEE, Kalé has received the 2002 ACM Gordon Bell Prize (with his co-authors) and the 2012 IEEE Computer Society Sidney Fernbach Award (with Klaus Schulten). He and his team won the HPC Challenge award at Supercomputing 2011, for their entry based on Charm++. He also co-edited the book Parallel Science and Engineering Applications: The Charm++ Approach.

Josep Torrellas: The Saburo Muroga Professorship in Computer Science

A pioneer in parallel computer architectures, CS Professor Josep Torrellas has made important contributions to shared-memory multiprocessor design, including in cache hierarchies, coherence protocols, synchronization, consistency models, and thread-level speculation. These contributions make it easier to program parallel computers while enhancing their performance.

Josep Torrellas
Josep Torrellas

About the Saburo Muroga Professorship

Established by Douglas B. MacGregor (MS CS '80) to provide significant recognition of an outstanding faculty member, the Saburo Muroga Professorship in Computer Science honors the late Professor Muroga for his service and dedication to students as manifested through exemplary teaching and guidance. After earning a PhD from Kyoto University in 1990, MacGregor became an executive with Data General Corporation and Dell, formed a multi-million dollar joint venture with Matsushita, and served on the faculty of the Harvard Business School.

In addition, his work has improved the energy efficiency of multiprocessor architectures. He has devised techniques to handle process variation and wear-out, and to reduce the power consumption of extreme-scale computer systems.

Torrellas has contributed to several leading industry–government research projects in novel parallel computer architectures. These include the Illinois Aggressive Cache Only Memory Architecture (IACOMA), which was one of the 10 Point-Design Studies funded by the federal government in the 1990s to accelerate the development of peta-scale supercomputers. He also led the DARPA-funded M3T Polymorphic Computer Architecture and codirected the NSF-funded FlexRAM Intelligent Memory project.

Torrellas was one of the principal investigators in the DARPA-funded IBM PERCS multiprocessor project, which led to the initial design of the Blue Waters supercomputer. He was also a co-leader of the DARPA- and DOE-funded Intel Runnemede multiprocessor, a 1000-core extreme-scale chip developed under the Ubiquitous High Performance Computing program. Before that, Torrellas contributed to the Stanford DASH and Illinois Cedar experimental multiprocessors.

Today, Torrellas is the director of the Center for Programmable Extreme-Scale Computing, which focuses on developing programmable, high performance, and very energy-efficient computers. He has been the director of the Illinois-Intel Parallelism Center (I2PC), whose aim was to promote parallel computing. In that center, he has worked with Intel to prototype deterministic replay multiprocessor hardware and developed the novel Bulk Multicore concept.

Torrellas is a Fellow of the AAAS, ACM, and IEEE. He has received the 2015 IEEE Computer Society Technical Achievement Award and the 2012 ICCD High-Impact Paper Award, among other awards and honors. He has graduated more than 35 doctoral students, of which 13 are faculty in top CS and ECE departments across the nation.


Share this story

This story was published November 14, 2016.