The AI hardware accelerator market has exploded with players in recent years, the result of over $10b startup investor funding, with still new players popping up out of stealth. In this Q&A session between HPCwire's Tiffany Trader and Kisaco Research's Michael Azoff we examine how the AI chip market has segmented itself, address the differences and commonalities between datacenters and HPC environments, as well as other segments. And we ask how sustainable is this market and what can we expect in the near and longer terms. The conclusions of this webinar are drawn from Kisaco Research’s recent vendor comparison reports on AI accelerator vendors across the HPC, Edge/IoT & Automotive sectors.
You will learn:
• How to compare AI chips. Which are the best metrics?
• How the AI chips differ, in architecture, in combination of software and hardware, in data pre-processing.
• The variables to consider: training, inferencing, number precision, power consumption and energy supply, size of chip, cost, and ease of use for developers.