Buch, Englisch, Format (B × H): 155 mm x 235 mm
Reihe: Hardware and Software Technologies for Artificial Intelligence Revolution
From Semiconductor Scaling to Photonic and Quantum Computing
Buch, Englisch, Format (B × H): 155 mm x 235 mm
Reihe: Hardware and Software Technologies for Artificial Intelligence Revolution
ISBN: 978-3-032-38949-7
Verlag: Springer
The data center has become the unit of computation for artificial intelligence, and the integrated circuit has become its limiting factor. AI models scale on one curve, while transistors, interconnects, memory, and power delivery scale on entirely different ones. The widening distance between them now decides what AI can afford to do. Most of the energy an AI data center consumes goes into moving data rather than performing arithmetic on it. That single fact reorders everything downstream: which architectures win, where the capital goes, and how much further CMOS can carry the industry. This book follows that logic from the transistor to the rack, beginning with how CPU, GPU, TPU, and custom accelerator architectures differ. It then traces the chip developments from Nvidia, AMD, Arm, and Intel now serving the AI ecosystem of OpenAI, Meta, Microsoft, Google, Apple, and xAI. Later chapters work through the crossover from copper to light that governs interconnect design, then rise to system-level architecture and data center economics. The closing chapters look past CMOS to hyperdimensional and neuromorphic computing, silicon photonics, advanced materials, and quantum processors, and ask which will run real workloads, and when. Written by leaders from industry and academia, this is a unified playbook for architects, system designers, product managers, and investors who need to translate semiconductor trends into AI performance, cost, and energy efficiency.
Zielgruppe
Professional/practitioner
Autoren/Hrsg.
Fachgebiete
Weitere Infos & Material
A Brief Introduction to AI.- Evolution of Silicon Computing Chips for the AI Era.- Mapping Neural Networks and Transformers to AI Hardware from an Energy Consumption Perspective.- Silicon Photonics and Connectivity for Data Centers.- The NeoCloud Revolution: Storage-Centric Acceleration and Emerging AI Architecture.- System-Level Hardware Evolution for Artificial Intelligence.- Hyperdimensional Computing for Energy-Efficient AI Hardware: From Hypervector Operations to Memory-Centric IC Implementations.- Architectural Roles and Integration of Quantum, Photonic, and Neuromorphic Computing.- Integrating Quantum Computing into AI Data Center Infrastructure.- Beyond Classical Limits: Quantum Computing as a Computational Engine for AI Workloads.




