Performance space modeling for hierarchical synthesis of analog integrated circuits

Gielen, Georges; McConaghy, Trent; Eeckelaert, Tom

doi:10.1145/1065579.1065809

Cited by 40 publications

(10 citation statements)

References 53 publications

(39 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…También se han publicado desarrollos de estrategias de optimización específicamente dirigidas a los circuitos MCML (Müller-Gritschneder, 2005;Gielen, 2005), pero únicamente con base en métodos deterministas (seguidores de gradiente), que tienen la desventaja de ser proclives a quedar atrapados en mínimos locales.…”

Section: Algoritmo Genético Para Optimizaciones Multiobjetivo De Circunclassified

Diseño de una biblioteca de compuertas MCML utilizando un algoritmo genético y optimización multiobjetivo

Pereira-Arroyo¹,

Chacón-Rodríguez²

2014

View full text Add to dashboard Cite

En este documento se alude al problema de dimensionamiento de circuitos MCML (MOS Current Mode Logic). Se introduce el Frente de Pareto como una herramienta de análisis útil para explorar el espacio de diseño de las distintas compuertas que conforman nuestra biblioteca básica MCML. Un algoritmo genético (GA) es implementado para detectar automáticamente este frente, en un proceso que busca eficientemente las parametrizaciones óptimas del diseño y sus correspondientes valores en un espacio de aptitudes. Las mediciones del consumo de potencia, el retardo de propagación y el rango del voltaje de salida se usan como funciones de aptitud, puesto que el problema se trata como una tarea de optimización multiobjetiva. Finalmente, se presentan los resultados de las simulaciones postlayout, usando la tecnología de fabricación AMS 0,35 μm.

show abstract

Section: Algoritmo Genético Para Optimizaciones Multiobjetivo De Circunclassified

Diseño de una biblioteca de compuertas MCML utilizando un algoritmo genético y optimización multiobjetivo

Pereira-Arroyo¹,

Chacón-Rodríguez²

2014

View full text Add to dashboard Cite

show abstract

“…We want to recast the problem of filter approximation as Performance Space Modeling [3]. Consider that a designer wants a filter approximation with a given Amax, Amin, ws, wp specification and objectives of maximizing phase linearity and minimizing peak-overshoot.…”

Section: Future Vision and Preliminary Experimentsmentioning

confidence: 99%

Filter approximation using explicit time and frequency domain specifications

Aggarwal

Jin

O’Reilly

2006

Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

We demonstrate that enhanced particle swarm optimization (PSO) can be successfully used to evolve high performance filter approximations. These evolved approximations use sets of quantitative specifications which conventional analytically derived approximations can not directly employ. The conventional derivations use only a subset of the quantitative specifications in their algorithm and the remaining specifications are side-effect results of the algorithm. Thus, with enhanced PSO, instead of a filter designer having access to a limited set of " specification knobs" that directly and indirectly achieve performance, a designer has a "knob" for each specification that consequently drives the approximation to desired performance. Our future vision to do performance space modeling for filters is also discussed.

show abstract

“…This algorithm may use lookup tables [1], radial basis functions (RBF) [2], artificial neural networks (ANN) [3,4] and its derivations such as fuzzy logic (FL) [5] and neural-fuzzy network (NF) [6], and regression [7,8]. Model generators can also be categorized into the black, grey and white box approaches, depending on the level of existing knowledge of the system's structure and parameters Unfortunately AMG may produce high order models of excessive complexity [9 -11], in which case model order reduction (MOR) techniques may be required [12]. A recent paper [13] discusses MOR techniques with respect to various model types, and in a variety of contexts: LTI MOR [14], LTV MOR [15,16], weakly nonlinear methods including polynomial-based [17,18], trajectory piecewise linear (TPWL) [19], and piecewise polynomial (PWP) [20].…”

Section: Introductionmentioning

confidence: 99%

A novel approach for automated model generation

Xia

Bell

Wilkinson

2008

2008 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

Automatic approaches for macromodel generation from SPICE-level descriptions have been of great interest over the last few years for the design of large, complex mixsignal SoCs (system-on-chips) and SiPs (system-in-packages). In this paper a novel approach termed multiple model generation system (MMGS) is developed for extracting either single-input single-output (SISO) or multiple-input singleoutput (MISO) macromodels from a SPICE netlist. This model generation process detects nonlinearity through variations in output error. Examples of the application of MMGS are presented for simple two-input systems incorporating a two-stage CMOS operational amplifier (op amp).I.

show abstract

Performance space modeling for hierarchical synthesis of analog integrated circuits

Cited by 40 publications

References 53 publications

Diseño de una biblioteca de compuertas MCML utilizando un algoritmo genético y optimización multiobjetivo

Diseño de una biblioteca de compuertas MCML utilizando un algoritmo genético y optimización multiobjetivo

Filter approximation using explicit time and frequency domain specifications

A novel approach for automated model generation

Contact Info

Product

Resources

About