What's New

Feature Article

Feature Article

Big Changes for Mainstream Chip Architectures

Using AI and non-von-Neumann architectures to continue exponential improvement in system performance

This special report from Semiconductor Engineering presents an overview of innovations in processor and memory architectures from several companies, intended to continue enhanced system performance over the next decade. Some of these innovations were presented at the Hot Chips Symposium held recently in Silicon Valley.

The new focus is not so much on process development, as on redesigning chip architectures and improving chip packaging, in order to optimize performance for particular applications. For example, edge devices will need to pre-process massive amounts of data, rather than sending all of this to the cloud. Furthermore, artificial intelligence built on neural networks will be distributed across processors and memories to enable learning to optimize scheduling and assignment among multiple devices on and off the chips. This will further deviate from the classic separation of logic and memory as in the Von Neumann architecture.

In this way, chip manufacturers anticipate that they will be able to double system performance every two years, even while the classic process-based Moore’s Law enhancements have mostly saturated.

Technology Spotlight

Technology Spotlight

The Future of Computing: A Conversation with John Hennessy

Dr. John Hennessy is a leading computer scientist who was the President of Stanford University and is now the Chairman of Alphabet. In May 2018, Dr. Hennessy presented a keynote talk at the Google I/O developer conference. He spoke about the future of computing in the context of the ending of Moore’s Law and the onset of artificial intelligence. Energy efficiency has become critical, and is limiting the performance of advanced processors. He focused on the need for “domain-specific architectures” and “domain-specific languages” that will enable the design and programming of special-purpose processors optimized for different applications. This differs from the general-purpose processors of earlier generations, and will enable accelerated performance without faster transistors. This will require closer integration of hardware and software with applications throughout the design phase, preferably including consideration of cybersecurity. Artificial intelligence and machine learning (AI/ML) based on neural networks are finally making major impacts on system performance.

Current CMOS technology should be sufficient for the next decade, but other technologies (neuromorphic, quantum) may be needed for further progress.

Watch The Future of Computing video

For other videos from Google I/O 2018, see here