David and Marilyn Karlgaard Endowed Chair Professor of ECE
Director, HPCAT Laboratory
Editor-in-Chief, IEEE Transactions on Computers
Associate Editor, IEEE Transactions on Cloud Computing
Associate Editor, IEEE Transactions on Sustainable Computing
Congratulations to Prof. Ahmed Louri, who was selected to receive the GWU 2021 Office of the Vice-President for Research (OVPR), Distinguished Researcher Award. The OVPR Distinguished Researcher Award recognizes GWU faculty members who have made significant contributions in research and scholarship to the university and society. The award is given annually to just one GWU faculty member.
Prof. Ahmed Louri is the recipient of the IEEE Computer Society 2020 Edward J. McCluskey Technical Achievement Award, “for pioneering contributions to the solution of on-chip and off-chip communication problems for parallel computing and manycore architectures.” (Award Video)
We rely on computing in the design of systems for energy, transportation, finance, education, health, defense, entertainment, and overall wellness. However, today's computing systems are facing major challenges both at the technology and application levels. At the technology level, traditional scaling of device sizes has slowed down and the reduction of cost per transistor is plateauing, making it increasingly difficult to extract more computer performance by employing more transistors on-chip. Power limits and reduced semiconductor reliability are making device scaling more difficult – if not impossible – to leverage for performance in the future and across all platforms, including mobile, embedded systems, laptops, servers, and datacenters. Simultaneously, at the application level, we are entering a new computing era that calls for a migration from an algorithm computing world to a learning-based, data-intensive computing paradigm in which human capabilities are scaled and magnified. To meet the ever-increasing computing needs and to overcome power density limitations, the computing industry has embraced parallelism (parallel computing) as the only method for improving computer performance. Today, computing systems are being designed with tens to hundreds of computing cores integrated into a single chip and hundreds to thousands of computing servers based on these chips are connected in datacenters and supercomputers. However, power consumption remains a significant design problem, and such highly parallel systems still face major challenges in terms of energy efficiency, performance, and reliability.
Professor Louri and his team investigate novel parallel computer architectures and technologies which deliver high reliability, high performance, and energy-efficient solutions to important application domains and societal needs. The research has far-reaching impacts on the computing industry and society at large. Current research topics include: (1) the use of machine learning techniques for designing energy-efficient, reliable multicore architectures, (2) scalable accelerator-rich reconfigurable heterogeneous architectures, (3) emerging interconnect technologies (photonic, wireless, RF, hybrid) for network-on-chips (NoCs) & embedded systems, (4) future parallel computing models and architectures including Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), near data computing, approximate computing, and (5) cloud and edge computing.
The award is given for outstanding and innovative contributions to the fields of computer and information science and engineering or computer technology, usually within the past 10, and not more than 15, years. (More)
National Science Foundation grant is for the project "Neural-Network-based Stochastic Computing Architectures with applications to Machine Learning". This award brings the total amount of federal funds for Prof. Louri for this research topic to $4 million over the past three years.
HPCAT has three papers accepted in 27th International Symposium on High-Performance Computer Architecture (HPCA'21).
1. CSCNN: Algorithm-hardware Co-design for CNN Accelerators using Centrosymmetric Filters
2.GCNAX: A Flexible and Energy-efficient Accelerator for Graph Convolutional Neural Networks
3.Adapt-NoC: A Flexible Network-on-Chip Design for Heterogeneous Manycore Architecture