Enhanced AEC Qualification of Groudbreaking 5nm AI Processor with Liquid Cooling for Level 4 Autonomous Driving

4 views
Download
  • Share
+0
Create Account or Sign In to post comments
#REPP #electronics packaging #photonics packaging

(30:25 + Q&A) Dr. Fen Chen, Nvidia.
The rapid adoption of artificial intelligence (AI) chips in autonomous vehicles (AV), driven by the need for complex computations, necessitates robust thermal management and stringent reliability testing. Liquid cooling has emerged as a critical technology to address the thermal challenges posed by high-performance AI processors. However, the existing AEC- Q100 qualification standard, developed for air-cooled legacy chips, falls short of addressing the unique requirements of modern AV AI processors. This paper proposes an enhanced AEC-Q100 qualification methodology tailored to a custom 5nm AI processor, highlighting the need to revise the standard to consider factors such as mission profiles, cooling mechanisms, dynamic load testing, and electromagnetic compatibility. The successful qualification of the AI processor underscores the importance of evolving industry standards to ensure the reliability and performance of next-generation high-performance processors that are pivotal for the advancement of autonomous vehicle technology.  (more)
Bio: Dr. Fen Chen is a recognized expert in Reliability, Availability, and Serviceability (RAS) engineering, with a Ph.D. in Electrical Engineering from University of Delaware, and extensive experience across high-tech industries. He specializes in ensuring the robust performance and uptime of critical systems and chip components. Currently, he serves as the RAS Subject Matter Expert for data center AI products at Nvidia. Previously, at Cruise, Dr. Chen led the reliability team, overseeing the validation of liquid-cooled compute systems and AI processors essential for autonomous vehicle operations. His work focused on maximizing reliability and serviceability to enhance the safety and efficiency of complex systems. Throughout his career, he has held key reliability roles at Apple and IBM Microelectronics and has chaired the JEDEC JC14.2 WLR committee. A prolific innovator and thought leader, Dr. Chen holds over 50 patents and has authored more than 60 publications.

Edited videos and slides from most of the REPP talks are available at https://attend.ieee.org/repp/?page_id=2110  Join our Dlist to hear about future REPPs: https://attend.ieee.org/repp/?page_id=361 
Organized by the IEEE Silicon Valley chapter of the Electronics Packaging Society: https://ieee.org/scveps 

(30:25 + Q&A) Fen Chen, Nvidia.
The rapid adoption of artificial intelligence (AI) chips in autonomous vehicles (AV), driven by the need for complex computations, necessitates robust thermal management and stringent reliability testing. Liquid cooling has emerged as a critical technology to address the thermal challenges posed by high-performance AI processors. However, the existing AEC- Q100 qualification standards… (more)

Speakers in this video

Advertisment

Advertisment