Ascend AI Processor Architecture and Programming

Artificial intelligence (AI) research is rapidly developing, with AI applications predicted to profoundly affect all areas of industry in the near future. Consequently, there is an ever-growing demand for processor chips that support AI research and applications. Currently, chips offering this support are limited to GPUs, TPUs, and the Huawei Ascend AI chip, with limited detailed information on these processors.

The Ascend AI chip is a powerful and versatile AI processor that will see extensive deployment across the globe. AI Processor Architecture and Programming offers in-depth AI applications using Huawei’s Ascend chip, presenting and analyzing the unique performance and attributes of this processor. The title introduces the fundamental theory of AI, and the software and hardware architecture of the Ascend AI processor, as well as related tools and programming technology, and typical application cases. The book demonstrates internal software and hardware design principles, system tools and programming techniques for the processor. It also lays out those elements of AI programming technology needed by researchers developing AI applications. The book consists in six chapters, covering the theoretical fundamentals of AI and deep learning; the state of the industry, including the current state of Neural Network Processors, deep learning frameworks, and a deep learning compilation framework; the hardware architecture of the Ascend AI processor; programming methods and practices for developing the processor, and finally; detailed case studies on data and algorithms for AI. This title presents the hardware and software architecture of the Huawei Ascend AI processor, supporting the future development and deployment of powerful AI applications.