This book provides a systematic and unified methodology, including basic principles and reusable processes, for dynamic memory management (DMM) in embedded systems. The authors describe in detail how to design and optimize the use of dynamic memory in modern, multimedia and network applications, targeting the latest generation of portable embedded systems, such as smartphones. Coverage includes a variety of design and optimization topics in electronic design automation of DMM, from high-level software optimization to microarchitecture-level hardware support. The authors describe the design of multi-layer dynamic data structures for the final memory hierarchy layers of the target portable embedded systems and how to create a low-fragmentation, cost-efficient, dynamic memory management subsystem out of configurable components for the particular memory allocation and de-allocation patterns for each type of application. The design methodology described in this book is based on propagating constraints among design decisions from multiple abstraction levels (both hardware and software) and customizing DMM according to application-specific data access and storage behaviors.

This book contains extended and revised versions of the best papers presented at the 18th IFIP WG 10.5/IEEE International Conference on Very Large Scale Integration, VLSI-SoC 2010, held in Madrid, Spain, in September 2010. The 14 papers included in the book were carefully reviewed and selected from the 52 full papers presented at the conference. The papers cover a wide variety of excellence in VLSI technology and advanced research. They address the current trend toward increasing chip integration and technology process advancements bringing about stimulating new challenges both at the physical and system-design levels, as well as in the test of theses systems.

Industrial Applications of Machine Learning shows how machine learning can be applied to address real-world problems in the fourth industrial revolution, and provides the required knowledge and tools to empower readers to build their own solutions based on theory and practice. The book introduces the fourth industrial revolution and its current impact on organizations and society. It explores machine learning fundamentals, and includes four case studies that address a real-world problem in the manufacturing or logistics domains, and approaches machine learning solutions from an application-oriented point of view. The book should be of special interest to researchers interested in real-world industrial problems. Features Describes the opportunities, challenges, issues, and trends offered by the fourth industrial revolution Provides a user-friendly introduction to machine learning with examples of cutting-edge applications in different industrial sectors Includes four case studies addressing real-world industrial problems solved with machine learning techniques A dedicated website for the book contains the datasets of the case studies for the reader's reproduction, enabling the groundwork for future problem-solving Uses of three of the most widespread software and programming languages within the engineering and data science communities, namely R, Python, and Weka

This book provides a systematic and unified methodology, including basic principles and reusable processes, for dynamic memory management (DMM) in embedded systems. The authors describe in detail how to design and optimize the use of dynamic memory in modern, multimedia and network applications, targeting the latest generation of portable embedded systems, such as smartphones. Coverage includes a variety of design and optimization topics in electronic design automation of DMM, from high-level software optimization to microarchitecture-level hardware support. The authors describe the design of multi-layer dynamic data structures for the final memory hierarchy layers of the target portable embedded systems and how to create a low-fragmentation, cost-efficient, dynamic memory management subsystem out of configurable components for the particular memory allocation and de-allocation patterns for each type of application. The design methodology described in this book is based on propagating constraints among design decisions from multiple abstraction levels (both hardware and software) and customizing DMM according to application-specific data access and storage behaviors.

Vertically-integrated 3D Multiprocessor Systems-on-Chip (3D MPSoCs) provide the means to continue integrating more functionality within a unit area while enhancing manufacturing yields and runtime performance. However, 3D MPSoCs are subject to amplified thermal challenges that undermine the corresponding reliability. To address these issues, several advanced cooling technologies alongside temperature-aware design-time optimizations and run-time management schemes have been proposed. This book surveys recent advances in temperature-aware 3D MPSoC considerations. It explores the recent advanced cooling strategies, thermal modeling frameworks, design-time optimizations, and run-time thermal management schemes that are primarily targeted for 3D MPSoCs. As such, it provides a global perspective, highlighting the advancements and drawbacks of the recent state-of-the-art. While the primary focus is on existing methodologies and techniques, it also provides some insights on possible future directions related to this research discipline - temperature-aware design and management. This is an ideal primer for researchers and practitioners working in this area.