This is a book that looks into the possibilities and potential benefits of integrating UML and SoC. UML is the unified modeling language, which is actually a family of visual notations derived from a number of previously existing modeling languages from the software engineering domain. Because UML unified different modeling approaches, which include several structural and behavioral views, the language achieved a level of expressiveness that allowed for its use in application domains other than software design. SoC means system-on-a-chip, denoting the current method for designing integrated circuits that supports the integration, in a single silicon die, of several electronic subsystems previously connected as discrete components in a printed circuit board. SoC design is a very heterogeneous discipline, and it tackles the sheer complexity of designing integrated circuits by relying on a number of abstraction layers supported by design languages and models.
The rationale behind this book is to analyze the potential of UML to be used as a design language within the SoC design flow, and to have a preview of the benefits that such increased abstraction can bring to SoC designers. This is an emerging topic, so there is no major approach to be followed and the state of the art is still diffuse. Several different--and mostly unrelated--approaches were included in this book, chosen from the major contributions to the UML for SoC Design Workshop 2004 (a fringe event to the IEEE/ACM Design Automation Conference). The contributions come from industry and academic researchers from Europe, Asia, and the Americas, and are each presented in a separate chapter.
The book opens with an introduction to the topic by the editors. It includes a number of motivations for using UML in SoC design, an overview of the main features of UML2, and some account of how UML2 can fit into the SoC design flow.
The second chapter advocates for the use of a restricted subset of UML (the so-called executable UML), which adopts only three of the UML notations--class diagram, statecharts, and actions--materialized through an action language. By relying on these constructs, developers should be able to build models that can be compiled into executable code tailored to a given execution platform.
Chapter 3 analyzes the capabilities of UML to capture system dynamics using different models of computation. It investigates the possibility of joint usage of state machines and activity diagrams, the former being used to capture control flow-dominated subsystems while the latter targets the modeling of subsystems with synchronous dataflow behavior.
Chapters 4 through 8 present approaches to the integration of UML into existing SoC design flows. Chapter 5 discusses “Hardware/Software Codesign of Reconfigurable Architectures Using UML.” Chapters 6, 7, and 8 integrate UML with flows based on the SystemC language, while chapter 4 addresses Handel-C code generation targeting an execution platform based on programmable logic. Chapter 7 presents a case study on the modeling of instruction set architectures with UML.
Chapter 9 provides a comparison between UML and languages based on functional blocks. The last chapter presents a method for optimizing power consumption in communication systems. The method is based on UML diagrams annotated with power consumption figures extracted from a simulation model.
This book is mainly targeted to SoC researchers and tool developers. It requires good familiarity with SoC design flows, languages, and tools. Some background in UML is desirable, as the review of UML2 features in chapter 1 is very brief and assumes that the reader is familiar with features of earlier versions of UML or the modeling languages that preceded it. Except for the introduction, the order of the chapters is irrelevant, so readers may go directly to their topics of interest. Unfortunately, the editors didn’t provide a summary of each of the chapters in the introduction, so readers should consider reading all of the chapter abstracts first.
