Lots of Chip Designing on Tap: From Chiplets to Systems
What you’ll learn:
- How chiplets are addressing AI in the data center.
- Chiplet integration styles.
- Chiplet technology trends.
Artificial-intelligence and machine-learning (AI/ML) applications are pushing the performance envelope to the point where chips can’t get any larger on a single die. Enter chiplets.
Chiplets can bring together different die technologies together in a single package to address AI/ML applications (see figure). This is already being done but in a proprietary fashion by large OEMs like NVIDIA. However, standards like UCIe will enable others to provide and utilize chiplets.
Dr. Jawad Nasrullah, CEO at Palo Alto Electron, gave a presentation (watch the video above) that takes a look at why we need chiplets, where we are now, and what the biggest challenge will be for designers. Watch the video to get the answers to those all-important questions.