 Название: Machine Learning-based Design and Optimization of High-Speed Circuits
Автор: Vazgen Melikyan
Издательство: Springer
Год: 2024
Страниц: 351
Язык: английский
Формат: pdf (true)
Размер: 22.1 MB
The book systematically expounds the main results obtained by the author in the field of design and optimization of high-speed integrated circuits (ICs) and their standard blocks (heterogeneous ICs, analog-to-digital and digital-to-analog converters, input/output cells, etc.) operating in non-standard conditions (deviations of technological process parameters, supply voltage, ambient temperature, etc.). The proposed methods are based on machine learning and consider effects of different external and internal destabilizing factors (radiation exposure, self-heating, nonideality of the power source, input signals, interconnects, power rails, etc.).
The main goals of most proposed methods and solutions of design and optimization of high-speed ICs are to improve important parameters and characteristics (performance, power consumption, occupied area on the die, transmitting and receiving data quality and accuracy) of circuits and reduce the effects of non-standard operating conditions and different types of destabilizing factors.
This book describes machine learning-based new principles, methods of design and optimization of high-speed integrated circuits, included in one electronic system, which can exchange information between each other up to 128/256/512 Gbps speed. The efficiency of methods has been proven and is described on the examples of practical designs. This will enable readers to use them in similar electronic system designs. The author demonstrates newly developed principles and methods to accelerate communication between ICs, working in non-standard operating conditions, considering signal deviation compensation with linearity self-calibration. The observed circuit types also include but are not limited to mixed-signal, high performance heterogeneous integrated circuits as well as digital cores.
The development of means of design and optimization of high-speed ICs has now become a decisive part in the process of IC design. The problems, solved during design and especially the optimization of high-speed ICs, require huge amount of computations because of enormous number of components and options of considering versions of designing circuits and their operating modes. From this point of view and taking into account current rise of Machine Learning (ML), connected with the occurrence of Big Data and more powerful computing resources in the last few years, ML methods and tools can be considered as most suitable during design and optimization of high-speed ICs. This monograph is devoted to description of the developed new principles, methods, and circuit solutions for design and optimization of high-speed ICs, based on ML. In the monograph different effective new principles, methods, solutions, and means of design and optimization of high-speed ICs are proposed.
The book is anticipated for scientists and engineers specializing in the field of IC design and optimization as well as for students and postgraduate students studying disciplines related to IC design.
Contents:
1. Means to Accelerate Transfer of Information Between Integrated Circuits
2. Design Methods of Integrated Circuits, Working Under Non-standard Operating Conditions
3. Signal Transmission Calibration Systems in Integrated Circuits
4. Methods to Improve Linearity of Signal’s Analog-to-Digital Conversion with Self-Calibration
5. Design of High-performance Heterogeneous Integrated Circuits
6. Design of Digital Integrated Circuits by Improving the Characteristics of Digital Cells
Скачать Machine Learning-based Design and Optimization of High-Speed Circuits
|
Навигация
Вход на сайт
Реклама
Machine Learning-based Design and Optimization of High-Speed Circuits 
