The Advanced Computing Systems (ACS) research program drives innovation in advanced computing through research on complex computational systems, from hardware architecture to applications. ACS is located within the bwtech@UMBC Research and Technology Park adjacent to the University of Maryland-Baltimore County (UMBC) campus. ACS has active research projects in neuromorphic and probabilistic computing, advanced modeling and simulation, energy efficiency, architectures, productivity, and resilience. We collaborate extensively with industry, academia, and national labs. Contact us.
We perform research on new innovative data-centric computer architectures that restore the balance between processor, memory, and IO and which distribute computation and data to minimize the power and performance implications of moving large amounts of data between isolated/remote processing and memory/IO systems.
High Performance Computing (HPC) demands the highest performance technology combined with intricate programming to deliver efficient and effective solutions to the most challenging and complex computational problems. HPC systems are growing increasingly complex. Faults that occur during runtime are harder and harder to diagnose. Silent data corruptions, where bits get flipped and answers change without being detected at runtime, are more and more common. We are researching technologies, algorithms and the feasibility of using probabilistic hardware to deliver 20x or more operations per Joule improvement to mission applications compared to commercial, non-probabilistic hardware. Accepting “faults”, willing to have imperfections, in exchange for more operations per joule allows users of HPC to get “good enough” answers to their problems far more quickly and efficiently than requiring “perfect” reliability
Research new computer architecture and programming models for data intensive problems which are difficult to solve on modern but conventional computer systems. Focus includes high performance tensor-based knowledge discovery, streaming analytics, and run time system research.
MODELING, SIMULATION AND EMULATION
The development of technology and tools to model and simulation proposed computer architectures and examine emerging components is essential to the creation of high performance computing (HPC) systems that balances performance with power consumption, reliability and cost. Modeling, Simulation and Emulation researchers actively guide and cultivate a community of researchers creating such capabilities.
The Neuromorphic Computation Research Program (NCRP) explores the benefits of neuromorphic computation for developing advanced machine learning frameworks and applying them to national security missions. NCRP researches advanced hardware architectures to identify the most power efficient methods of computing machine learning algorithms. Longer term research includes identifying how the brain computes information and sees if these concepts can be incorporated into hardware neuromorphic processor designs.
At almost all scales, from handheld devices to large supercomputers, modern computing systems are power limited. The amount of computing that can be performed on a chip, in a cabinet, or in a room is limited by power dissipation. Research the technologies and architectures of high performance computers to minimize power consumption while maximizing performance, thus improving the overall power efficiency of computer systems.