CPUsNews

AMD’s Zen Architecture was Designed by a Team of More than 2,400 from Austin, Colorado, and India

It’s no secret that without the advent of the Zen CPU architecture, AMD would have most likely gone bankrupt, or at the very least continued to remain in Intel’s shadows. One of the key architects to come up with the floorplan and basic design of this architecture was Jim Keller. However, he was only one of the several hundred (and later thousand) people who was involved in the development of Zen. In an interview with AnandTech, Keller shared some interesting behind-the-scene details regarding the design and development of the Bulldozer successor.

As per Keller, he led a team of up to 2,400 people including directors, senior fellows, and programmers encompassing multiple different teams. There was the SoC team, Fabric team, and IP teams, all working together from multiple different locations. The SoC team was based out of Austin and partly from India, while the floating-point cache design was done in Colorado. The Zen core frontend was also done in Austin, while the front-end of the scrapped Arm chip was done in Sunnyvale.

Some of the key people involved in the design of Zen were Suzanne Plummer and Steve Hale who worked on the front-end, while the rest was handled by Mike Clark, Leslie Barnes, Jay Fleischman, Harry Fair, and a few others. Most of these folks are still working at AMD. Therefore, as you can see a lot of people were behind Zen, making Keller one of its uncles.

When Keller first took the role, it took a fair bit of convincing to get the entire team on board, as a chiplet-based design connected using “glue” was a radical new approach with its own risks. The team grew from a relatively modest 500 to an impressive 2,400 strong by the time the design was done and Keller decided to move on.

Areej

Computer Engineering dropout (3 years), writer, journalist, and amateur poet. I started my first technology blog, Techquila while in college to address my hardware passion. Although largely successful, it was a classic example of too many people trying out multiple different things but getting nothing done. Left in late 2019 and been working on Hardware Times ever since.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top button