Overcoming nanoscale variations through statistical error compensation
Gao, Tianqi
Loading…
Permalink
https://hdl.handle.net/2142/78602
Description
Title
Overcoming nanoscale variations through statistical error compensation
Author(s)
Gao, Tianqi
Issue Date
2015-04-07
Department of Study
Electrical & Computer Eng
Discipline
Electrical & Computer Engr
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
variation-tolerant techniques
power-efficient design
emerging technologies
error compensation
Abstract
Increasingly severe parameter variations that are observed in advanced nanoscale technologies create great obstacles in designing high-performance, next-generation digital integrated circuits (ICs). Conventional design principles impose increased design margins in power supply, device sizing, and operating frequency, leading to overly conservative designs which prevent the realization of potential benefits from nanotechnology advances. In response, robust digital circuit design techniques have been developed to overcome processing non-idealities. Statistical error compensation (SEC) is a class of system-level, communication-inspired techniques for designing energy efficient and robust systems. In this thesis, stochastic sensor network on chip (SSNOC), a known SEC technique, is applied to a computational kernel implemented with carbon nanotube field-effect transistors (CNFETs). With the aid of a well developed CNFET delay distribution modeling method, circuit simulations show up to 90× improvement of the SSNOC-based design in the circuit yield over the conventional design. The results verify the robustness of an SEC-based design under CNFET-specific variations. The error resiliency of SEC allows CNFET circuits to operate with reduced design margins under relaxed processing requirements, while concurrently maintaining the desired application-level performance.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
