XPS: FULL: FP: Design and Synthesis of New Energy-efficient Self-healing Computing Electronics with Real-time Configurability

Project: Research project

Description

The energy consumption and material cost for modern data centric computing system have been rapidly growing through the last decade. Unfortunately, a significant amount of the energy and material expense has been wasted to create enough design tolerance to manufacturing defects, variability and reliability degradation, all of which have been exacerbated at each generation of semiconductor technology. As a result, the current worst-case design methodology with a large design margin for computer hardware has created a huge burden to our economy and environmental system due to the inefficient use of energy and material. As the technology scaling has become prohibitively expensive, a novel design methodology with real-time reconfigurability to tolerate large variability from both manufacturing and user demand is highly desirable. Unfortunately, conventional CMOS technology has not provided us such a capability because once an integrated circuit is fabricated, it can no longer be altered leading to a large material and energy overhead in exchange of robustness and reliability. However, the very recent development of non-volatile memresistive device has opened the door for a whole new design methodology with real-time reconfigurability.

This work explores a design and synthesis methodology to create a new energy and cost efficient computer system with “self-healing” capability. By intelligently deploying novel self-healing circuit building blocks into modern microprocessors, the new computer system is able to perform real-time reconfiguration to combat the performance loss due to process variation or aging effect. As a new design paradigm with real-time power and performance tradeoff is created from the proposed scheme, a novel design synthesis and automation process will be developed to fully materialize the self-healing capability. The proposed development will provide a significant improvement in energy efficiency as well as system robustness to the escalating variability and reliability issues in the modern semiconductor technology and thus lead to a new scalable design methodology for the next generation computing system.

Intellectual Merit Summary:
The proposed research creates a highly desirable self-healing computing system that has not been realized before. The proposed development consists of (1) novel circuit techniques using emerging memresistive device to bring an unprecedented reconfigurability to the existing computer electronics, (2) a new cross-layer systematic design methodology which maximize the benefits of the self-healing capability and provide real-time tunability, (3) integrated modeling and tooling support for digital and mixed-signal design using the reconfigurable emerging memresistive device.

Broader Impacts Summary:
This research will not only create a new design paradigm of next generation reconfigurable computing system but also will help reduce the development cost of emerging semiconductor technology and thus facilitate the creation of new technology and business. The cross-layer nature of the research project promotes a multi-disciplinary research methodology and trains students with broad knowledge base and out-of-box thinking.
StatusFinished
Effective start/end date9/1/158/31/19

Funding

  • National Science Foundation (CCF-1533656)

Fingerprint

Electronic equipment
X ray photoelectron spectroscopy
Semiconductor materials
Computer systems
Costs
Networks (circuits)
Computer hardware
Integrated circuits
Energy efficiency
Microprocessor chips
Energy utilization
Automation
Aging of materials
Students
Degradation
Defects