Название: Numbers and Computers, 3rd EditionАвтор: Ronald T. Kneusel
Издательство: Springer
Серия: Texts in Computer Science
Год: 2025
Страниц: 409
Язык: английский
Формат: pdf (true), epub
Размер: 36.1 MB
This textbook details the variety of number formats used by computers, thereby helping to ground readers in what can and cannot be represented accurately, especially by floating-point numbers.
The book's first part details standard representations of integers and floating-point numbers. The second explores other number representations, including the wide variety recently developed to support Artificial Intelligence (AI) and its demand for efficiency in representation to accommodate the ever-expanding scope of neural network models. Chapters describe each format, with examples in code (Python and C) and exercises. This new edition includes three new chapters on posits, AI number formats, and a collaborative experiment with an AI to generate novel number formats. Additionally, the existing text has been revised and edited and software updated as necessary to move from Python 2.7 to Python 3. The book’s software is available on GitHub.
If something is not easy to see in the code, it will be described in the text. Why C? Because C is a low-level language, close to the numbers we will be working with and because C is the grandfather of most common programming languages in current use including Python. In general, code will be offset from text and in a monospace font. For readers not familiar with C and/or Python there are a plethora of tutorials on the web and reference books by the bookcase.
At the end of each chapter are references for material presented in the chapter. Much can be learned by looking at these references. Almost by instinct, we tend to ignore sections like this as we are now programmed to ignore advertisements on web pages. In this former case, resist temptation; in the latter case, keep calm and carry on.
Topics and features:
Explores how computers use numbers to complete operations
Adds new chapters on posits and AI number formats
Includes exercises and examples that are code snippets in C or Python
Implements and tests new AI-designed number formats (as designed by GPT-4)
Provides thorough grounding on what can and cannot be represented accurately
Who Should Read This Book:
A textbook eminently suitable for undergraduates in Computer Science, the work also will appeal to software developers, engineers, scientists, AI experts, and anyone who programs for fun.
This book is for anyone who develops software including software engineers, scientists, computer science students, engineering students, and anyone who programs for fun.
If you are a software engineer you should already be familiar with many of the topics in this book, especially if you have been in the field for any length of time. Still, I urge you to press on, for perhaps you will find a gem or two which are new to you.
If are old enough, you will remember the days of Fortran and mainframes. If so, like the software engineers above, you are probably also familiar with the basics of how computers represent and operate on numbers, but, also like the software engineers above, you will likely find a gem or two of your own. Scientists in particular should be aware of the limitations and pitfalls associated with using floating point numbers since few things in science are restricted to integers.
Students need this book because it is essential to know what the computer is doing under the hood. After all, if you are going to make a career of computers, why would you not want to know how the machine works?
Скачать Numbers and Computers, 3rd Edition