Architecture of Computing System based on Chiplet

Area of Science:

• Computer Engineering
• Electrical Engineering
• Materials Science

Background:

• Computing systems are integral to modern technologies like medical diagnosis, climate prediction, and autonomous vehicles.
• The performance of computing systems directly impacts the intellectualization and capabilities of electronic equipment.
• Optimizing performance, efficiency, and cost in computing systems requires careful selection of architecture.

Purpose of the Study:

• To review Chiplet-based computing system architectures for high-performance applications.
• To provide guidance for designers on fabricating advanced computing systems.
• To analyze both computing and memory architectures within Chiplet-based systems.

Main Methods:

• Summarizing computing architectures for high-performance, mobile, and PC applications.
• Introducing mainstream and emerging non-volatile memory architectures for data storage and processing.
• Comparing and discussing key parameters of various memory types.

Main Results:

• A comprehensive overview of Chiplet-based computing architectures is presented.
• Analysis of different memory architectures, including their parameters and suitability for data handling.
• Identification of trade-offs between performance, efficiency, and cost in Chiplet designs.

Conclusions:

• Chiplet-based architectures offer a viable solution for high-performance computing systems.
• The choice of computing and memory architecture is critical for system optimization.
• Future perspectives on Chiplet-based computing systems are discussed, highlighting potential advancements.