Technology Spotlight - 2015
Low-Power Image Recognition Competition
Real-time automated image recognition is a key task in modern computing, and is increasingly being carried out on low-power mobile devices. As a way of developing interest in the field among both students and industry, Rebooting Computing is sponsoring an annual competition between teams to develop methods (optimizing both hardware and software) to achieve speed and accuracy together with low power. Prof. Yung-Hsiang Lu of Purdue University described the first competition and the winning teams at a recent meeting of the Embedded Vision Alliance, an industry group promoting the developing technology of Machine Vision.
For further information on LPIRC and the 2016 competition, see here.
For further information on the Embedded Vision Alliance, see here.
Software Architecture of HP’s new Memory Driven Computer, “The Machine”
HP’s radical new computer system, “The Machine”, will be based around enormous multi-Terabyte universal non-volatile memory that is directly addressable, accessed via fast optical links. Dr. Kim Keeton of HP Labs discusses the new software structures that will be required to take advantage of the new hardware architecture.
For further information on the architecture of The Machine, see here
Shaping the Future of Energy Efficient Supercomputing
The Hartree Centre is a UK Government-affiliated R&D Facility in High Performance Computing, located in Cheshire. A key focus is the enhancement of energy efficiency in terms of software, computer hardware, and data centers. Users must have tools to continuously monitor energy consumption at the program level, in order to properly assess tradeoffs.
For further information on the Hartree Centre Energy Efficiency Project, see here
Exascale Computing and Big Data
Prof. Daniel Reed of University of Iowa speaks briefly about how scientific supercomputing and Big Data/Cloud Computing were developed using completely different software approaches, but that success in the future requires that these two fields should work together.
A more complete discussion is available in an article by Reed and Dongarra in the July issue of Communications of the ACM.
Brain-Based Chips for Future Computing
Prof. Henry Markram of the Ecole Polytechnic Federale de Lausanne, Switzerland, and Director of the European Human Brain Project and the Blue Brain Project, presented a lecture on Neuromorphic Computing at the World Economic Forum at Davos, Switzerland.
For further information on the Human Brain Project see https://www.humanbrainproject.eu/
For further information on the Blue Brain Project see http://bluebrain.epfl.ch/
Future Computing in the US Army
Dr. John Pellegrino, Director of Computer Sciences at the US Army Research Laboratory, presents an overview of the future computer needs of the US military, as part of a special issue of Army Technology Magazine.
For further information on future computing research at the Army Research Laboratory, read more here.
Research at HP Labs: "The Machine"
CTO Martin Fink of HP recently presented a vision of a future energy efficient computer based on new memory devices, new computer architectures, and new software design. For example, this includes memristor-based memory technology, photonic data links, and a new open-source operating system.
For further information on "The Machine" from HP Labs, read more here.
The world's smallest computer has been developed at the University of Michigan, complete with power supply and input-output elements on the mm-scale. With a small solar cell for recharging the battery, a computer of this type may be ideal for autonomous sensors in the Internet of Things.
For further information on the Michigan Micro-Mote (M3) computer, read more here.
Automata Processing: A Novel Hardware/Software Approach for Massively Parallel Computing
Automata processing is a novel approach to parallel distributed memory and processing that can be much faster than conventional computing for certain problems, such as those involving searching and pattern matching of massive data sets. A technology of programmable silicon chips that can implement automata processing efficiently has recently been developed by Micron Technology.
For further information on Micron Automata Processor technology, read more here.
For a published paper that describes the technology in detail, see Paul Dlugosch et al., "An Efficient and Scalable Semiconductor Architecture for Parallel Automata Processing", IEEE Transactions on Parallel and Distributed Systems, available here.
The Next Consumer Technology Wave at Intel
At the recent International Consumer Electronics Show in Las Vegas, NV, Intel CEO Brian Krzanich presented several new computer technologies that will impact ubiquitous computing in consumer systems in the next several years. These include improvements in gesture control, wireless recharging, autonomous navigation for robots and drones, and a 1-cm computer for wearable and embedded applications.
For further information on Intel consumer electronics technology, read more here.