What is Computer Engineering Research?
Computer Engineering deals with the design, development, testing, and evaluation of components, systems, and networks. Research in Computer Engineering strives to acheive higher performance in the systems and components that are built as well as in the design process. At USC, research is focused in three main areas: Computer Architecture and Parallel Processing, Design and Analysis of Computer Networks and their Protocols, and VLSI Design and Computer Aided Design and Verification of VLSI
At USC, research is focused on three main areas:
1) Computer Architecture & Parallel Processing
2) Design & Analysis of Computer Networks and their Protocols
3) VLSI Design & Computer Aided Design and Verification of VLSI
Sample research projects include resilient and energy-efficient architecture design, design for test and testability, multiprocessor memory systems, internet computing, analysis and optimization of wireless ad hoc networks, green information technologies, adaptive routing for multi-core processors, and reconfigurable computing.
The faculty research is supported by NSF,SRC, DoD/DoE, and the industry.
The Computer Architecture group at USC is comprised of five faculty members. Four are IEEE Fellows and two are ACM fellows. We conduct research in widely diversified aspects of computer design and seek solution at every level of the hardware/software stack. The group has a long track record of major contributions in parallel architectures, reconfigurable architectures, memory systems, and interconnection networks, all critical areas of research in this new era of pervasive parallelism at all system levels and of power/energy design constraints.
At the micro-architecture level, current research focuses on understanding reliability challenges in deep sub-micro process technologies and providing (micro-) architectural solutions to address these concerns. Algorithmic optimizations are also explored to exploit state-of-the-art multi core and many core architectures (including GPUs) and to minimize peak power dissipation and energy consumption in FPGA-based application-specific architectures. At the system level, the group focuses on data center architecture and grid/cloud computing. Important research topics are trade-offs between power, performance and reliability, as well as security issues.
Faculty members in the Computer Architecture group are Murali Annavaram, Michel Dubois, Kai Hwang, Timothy Pinkston and Viktor Prasanna.
Working in collaboration with other colleagues in Electrical Engineering, Computer Science and the USC Information Sciences Institute, these researchers form one of the most active and visible networking research groups in the world with particular strengths in the design and analysis of protocols for next-generation wired and wireless networks that will form the future Internet, including Internet backbone infrastructure, peer-to-peer overlay networks, mesh networks, mobile ad-hoc and vehicular networks, sensor networks, intermittently-connected networks, underwater networks, cognitive radio networks, and mobile social networks.
One of the ongoing collaborative projects in this group pertains to maximizing throughput performance in wireless networks. This project involves the development of novel mathematical models of the capacity region of wireless networks and the design and testing of clean-slate dynamic routing and congestion control techniques that can be efficiently implemented in practical systems.
Another exciting new project aims to use mathematical tools from economics and game theory to develop a new class of robust spectrum sharing protocols for cognitive radio wireless networks that cannot be abused by selfish users. This project involves collaboration with researchers in the department of Economics at USC.
Researchers from this group are also starting to explore the emerging area of green networking and green information and communication infrastructure. Some projects in this area include design of sleep-scheduling algorithms for energy-efficient operation of cellular wireless networks, as well as novel algorithms and architectures for energy-efficient Internet backbone.
This group has substantial ties with industry, including in-depth collaborations with Bosch Research on wireless sensor networks, with General Motors CRD on the design of protocols and applications for vehicular ad hoc networks, with Sprint Labs on scheduling algorithms for WiMax systems, and with Nokia Research on energy-efficient sensing applications for mobile devices.
The group also participates in network research testbeds and academic network infrastructure nationally and internationally.
Faculty working in the research area of networks in the CENG group include Murali Annavaram, Rahul Jain, Bhaskar Krishnamachari, Viktor Prasanna, Konstantinos Psounis, Cauligi Raghavendra, and John Silvester.
The research philosophy of the group is to address revolutionary rather than evolutionary problems. Seldom are papers written where the main results deal with developing a heuristic that runs 10% faster than the previous best result, or does 5% better in power reduction or fault coverage. The faculty are well known for being the first to address several issue that later became mainstream to the community, such as crosstalk effects in design and test, CAD for asynchronous circuits, new paradigms in CAD such as integrating logical design with physical design and placement algorithms that are wire-length oriented, a variety of low-power design techniques, high-level synthesis of digital system, and efficient parallel mappings of algorithms to FPGA fabrics. Because of such innovative approaches, the Program has received many research awards. Faculty members have written several of the leading textbooks in the field of design, test and CAD. They interact with faculty in other fields, carrying out interdisciplinary research in areas such as building an artificial brain, improving the energy efficiency of data centers, oil field exploration and management, etc.
Faculty working in VLSI/CAD in the CENG group include Peter Beerel, Paul Bogdan, Melvin Breuer, Sandeep Gupta, Alice Parker, and Massoud Pedram.