AMD has historically been behind a generation of core IP for it’s APUs, why? SemiAccurate dug in to this complex topic and realized someone had really thought the big picture through down on Southwest Parkway.
The problem is simple, AMD comes out with core IP on a 12-ish month cadence, that means new CPU cores and GPU architectures. Currently this means Zen 3 cores are about to come out in the desktop Ryzen 4000 models while the mobile and desktop APU versions of the Ryzen 4000 line use Zen 2 and ‘last generation’ GPU cores.
If you ask pundits about this ‘problem’ you usually get one of two answers. First they will point out that once the IP is done it takes time to integrate it into other products with other IP on possibly differing schedules. Fair enough. They will also point out that AMD is short of engineering manpower and this situation will change once AMD has more money and headcount.
Both of these explanations sound plausible but are they what is really happening? Will we see the gap drop to a hypothetical three or six months between IP release and APU release? Will more cooks speed up the process? Once SemiAccurate examined the details it all clicked in to place and we came away with one conclusion, someone at AMD is very clever and thought this through in a big picture way, long term gain not short term headline grabs. Lets dive in to the details.
Note: The following is for professional and student level subscribers.
Disclosures: Charlie Demerjian and Stone Arch Networking Services, Inc. have no consulting relationships, investment relationships, or hold any investment positions with any of the companies mentioned in this report.