Keynote

David Greenhill

VP Silicon Engineering, Cerebras

Title:  Testing Wafer-Scale AI Compute

Biography:  David Greenhill has over 30 years’ experience designing high-performance computing systems. Currently he is VP of silicon engineering at Cerebras Systems Inc. His team develops all aspects of the wafer scale engine from architecture through to productization and test. Previously he has worked on the Inmos Transputer, Sun Microsystems server CPUs, Texas Instruments OMAP processors and Altera/Intel FPGAs. 

Abstract: The rapid growth of generative AI is reshaping the requirements for semiconductor systems. Trillion parameter large language models, AI reasoning, and Agentic AI push the need for very fast inference. The Cerebras Wafer Scale Engine (WSE) demonstrates a radically different approach: treating an entire silicon wafer as a single 46,225 mm² chip, integrating 4 trillion transistors and almost a million AI-optimized cores into one tightly coupled compute fabric.

The inherent advantage that 44GB of very high bandwidth on chip SRAM, tightly coupled to processing elements and low latency compute fabric enable the world’s fastest AI inference products. This keynote will explore the design, manufacture, and test of the WSE. We will explore the challenges Cerebras overcame to make this leap practical. Cerebras is the first company to solve the yield, redundancy, diagnosis, repair, burn-in, power delivery, and thermal management problems required to deliver wafer scale computing in volume production. Robust test methodologies and repair capabilities convert imperfect silicon into reliable compute capacity. Cerebras systems are now being deployed at scale by some of the world’s leading AI and cloud providers.