New open-access journal for novel devices and circuits for future computer systems. Papers are broadly read and highly cited, with rapid publication. Read the latest articles and submit your new research!
Computing in Science and Engineering has a feature article on development plans of the US Department of Energy for a supercomputer with 1018 operations per second, 100 times more powerful than today’s supercomputers.
Research into the structure and function of the brain is providing inspiration to a new generation of computer systems, which will enable faster and more efficient machine learning in future electronic devices.
Recent generations of microprocessors have critical overheating issues, requiring large parts of the circuits to be turned off when not in use, hence “Dark Silicon”. Careful design enables maximizing system performance under these constraints.
Recent developments in superconducting quantum computers by Google, D-Wave Systems, and IBM are discussed.
Reports by Technology Working Groups will be integrated into new technology roadmap to be issued by the end of 2017.
Contributions include both continued evolution and disruptive alternative technologies.
Stanford University, Feb. 7-8, 2017. Videos and slides of most keynote talks are available here.
University of Michigan researchers have developed ultra-small, ultra-low power (< 1mW) computers for application to mobile Internet of Things. The latest version incorporates neural networks with deep learning.
HiPEAC Vision Report advocates reinvention of computing, and predicts “cyber-physical entanglement of the physical and virtual worlds”.
Special issue reviews the past and includes a set of Outlook articles on the Future of Computing.
New White House Report on the Future of the US Semiconductor Industry, from President’s Council of Advisors on Science and Technology (PCAST).
More than 11,000 attendees broke records.
How Google used Neural Nets and Deep Learning to Transform Google Translate. Cover Story in New York Times Magazine.
A report was recently issued, based on the Arch2030 Workshop at ISCA2016. The end of classical scaling requires a wide range of new approaches across the entire computing stack from devices to software.
New research and new products demonstrate high-bandwidth optical transfer of data between processor chips.