Finding deterministic solution from underdetermined equation: large-scale performance modeling by least angle regression

Authors:
Xin Li
Affiliations:
Carnegie Mellon University, Pittsburgh, PA
Venue:
Proceedings of the 46th Annual Design Automation Conference
Year:
2009

Citing 8
Cited 3

Convex Optimization

Convex Optimization
Asymptotic probability extraction for non-normal distributions of circuit performance

Proceedings of the 2004 IEEE/ACM International conference on Computer-aided design
Projection-based performance modeling for inter/intra-die variations

ICCAD '05 Proceedings of the 2005 IEEE/ACM International conference on Computer-aided design
Performance-oriented statistical parameter reduction of parameterized systems via reduced rank regression

Proceedings of the 2006 IEEE/ACM international conference on Computer-aided design
Beyond low-order statistical response surfaces: latent variable regression for efficient, highly nonlinear fitting

Proceedings of the 44th annual Design Automation Conference
Principle Hessian direction based parameter reduction with process variation

Proceedings of the 2007 IEEE/ACM international conference on Computer-aided design
Statistical regression for efficient high-dimensional modeling of analog and mixed-signal performance variations

Proceedings of the 45th annual Design Automation Conference
Signal Recovery From Random Measurements Via Orthogonal Matching Pursuit

IEEE Transactions on Information Theory

Toward efficient large-scale performance modeling of integrated circuits via multi-mode/multi-corner sparse regression

Proceedings of the 47th Design Automation Conference
Finding deterministic solution from underdetermined equation: large-scale performance variability modeling of analog/RF circuits

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Statistical characterization of standard cells using design of experiments with response surface modeling

Proceedings of the 48th Design Automation Conference

Quantified Score

Hi-index	0.00

Visualization

Abstract

The aggressive scaling of IC technology results in high-dimensional, strongly-nonlinear performance variability that cannot be efficiently captured by traditional modeling techniques. In this paper, we adapt a novel L1-norm regularization method to address this modeling challenge. Our goal is to solve a large number of (e.g., 104~106) model coefficients from a small set of (e.g., 102~103) sampling points without over-fitting. This is facilitated by exploiting the underlying sparsity of model coefficients. Namely, although numerous basis functions are needed to span the high-dimensional, strongly-nonlinear variation space, only a few of them play an important role for a given performance of interest. An efficient algorithm of least angle regression (LAR) is applied to automatically select these important basis functions based on a limited number of simulation samples. Several circuit examples designed in a commercial 65nm process demonstrate that LAR achieves up to 25x speedup compared with the traditional least-squares fitting.