European Supercomputer Consortium Lays out Future Roadmap - Exascale systems targeted for 2023-2026.
Los Alamos Develops Binary-to-DNA Translator - New software package for long-term data storage and retrieval.
- IBM reports improved performance of high-speed, high-density memories - Spin-transfer-torque MRAM embedded in 14 nm circuits demonstrated with ns switching.
- 3D integration of heterogeneous chips for exascale processors - CEA (France) develops chiplet-level packaging with increased bandwidth and density.
- Graphene-Based Memristors for Improved Neural Networks - Penn State researchers demonstrate synapses with multiple synaptic weights
- Intel develops new cryptography protocols - Designed to be resistant to hacking by future quantum computers
- New memristor device acts like biological neuron - May provide basic element for neuromorphic computing
- Record speeds for AI inferencing - Nvidia GPU exceeds prior benchmark performance
- Analog Optical Computing Chip for Neural Networks - Carries out multiply-accumulate computations in silicon photonic chip for AI.
- US Government Announces New Research Centers in AI and QC - $1B in government labs and universities funded by National Science Foundation and Dept. of Energy.
- Newsletter of IEEE Electron Devices Society features new IRDS™ Roadmap - Provides history of earlier semiconductor roadmaps and an overview of the latest roadmap.
- US Government funds 3 academic-based institutes on quantum information - $75M over five years going to research based at the University of Colorado, University of California at Berkeley, and the University of Illinois.
- Chip Proposals Seek to Revive US Manufacturing - US Congress proposes aid to advance US semiconductor industry.
White Papers on the Future of Computing
The Computer Community Consortium is an advisory group of mostly US academic researchers representing the Computing Research Association. They sponsor Workshops and Task Forces, and issue periodic advisory White Papers to the US government research funding agencies. Recently, the CCC issued a series of Quadrennial White Papers covering a wide range of research and policy issues.
One of these is entitled “Advancing Computing’s Foundation of US Industry and Society,” by Thomas Conte (Georgia Institute of Technology), Ian Foster (University of Chicago), William Gropp (University of Illinois), and Mark Hill (University of Wisconsin).
The key point of this White Paper, as indicated in the Figure, is that there are at least 4 distinct approaches to future computing R&D, all of which need to be pursued in parallel. The first of these, “More Moore”, continues on traditional paths of enhancing the performance of CPUs, by continuing Moore’s Law scaling of transistor circuits. This will involve minimal disruption of the Computing Stack, particularly with respect to high-level software.
The second approach could add on sophisticated accelerators or co-processors to standard CPUs, in cloud-based data processing centers, in a way that is largely invisible to the end user. This would also be relatively non-disruptive.
But to achieve further enhancements in performance will require major changes in computing architecture. For example, efficient processing of Big Data will require moving the processing into the memory itself, in a way that may bypass the traditional CPU. This will require careful co-design of hardware and software, on the level of the end users.
Finally, the most disruptive changes are those associated with changing the traditional von-Neumann computing paradigm with completely different computing structures on all levels of the stack. This could include, for example, quantum computing, or neuromorphic computing, or analog computing.
2021 Technology Predictions of the IEEE Computer Society
The IEEE Computer Society revealed their annual top technology predictions for 2021. Download the full report.
The 2021 Technology Predictions focused in part on how COVID impacted the information environment, covering issues of AI, security, reliability, and core technologies.
For completeness, the IEEE CS team of experts also reviewed last year’s predictions to see whether they were correct. Not entirely, they admitted.
A video summary of the report is available in 3 brief YouTube videos, which play in sequence.
These are some of the many videos available for free from the IEEE Computer Society YouTube Channel. Check them out here.
In Pursuit of 1000X: Disruptive Research for the Next Decade of Computing
Keynote Video Lecture for Intel Labs Day, 3 December 2020
Coordinated by Dr. Richard Uhlig, Director of Intel Labs
In the past, the semiconductor industry depended on Moore’s Law to achieve exponential improvement in computing performance. But Moore’s Law is approaching its limits, so looking to the future, alternative approaches are needed to maintain growth. Intel Labs is carrying out advanced research into a variety of novel approaches in hardware and software that promise to enable improvements of 1000X or more in terms of speed, efficiency, or other performance metrics. Dr. Uhlig introduces Intel researchers who address the 5 following approaches:
(1) Integrated Photonics
(2) Neuromorphic Computing
(3) Quantum Computing
(4) Confidential Computing
(5) Machine Programming
An overview of the lecture is available at HPCwire.
The entire set of videos from Intel Labs Day is available at the Intel website.
- 2020 CCC Workshop on Physics and Engineering Issues in Reversible/Adiabatic Classical Computing
- Rebooting Computing Video Overview
- IEEE Future Directions
- IEEE Future Directions Blog
- Computing in Science and Engineering on the End of Moore's Law
- IEEE Journal of Exploratory Solid-State Computational Devices and Circuits (JXCDC)
- Arch2030 Workshop Report (PDF, 948 KB)
- Workshop on Neuromorphic Computing
- Workshop on Beyond CMOS Technology
- Update on National Strategic Computing Initiative (NSCI)
- RC White Paper on Nanocomputers
- IEEE Computer Magazine on Rebooting Computing