CPUsNews

AMD Zen 3D V-Cache Based Ryzen 6000/5000XT CPUs May be Available in December 2021 [Rumor]

Just the other day, it was discovered that AMD’s Milan-X processors featuring 3D stacked L3 cache (V-Cache) are all set to launch, with the retail OPNs and prices already finalized. Like Milan, Milan-X will be a full-fledged product stack with core counts ranging from 16 to 64. The existing Epyc processors already pack a massive L3 cache (up to 256MB). It’ll be interesting to see if AMD doubles it with Milan-X or simply change the topology.

Earlier today, Greymon55 stated that the 3D packaging equipment for Zen 3D already arrived in AMD’s China factories last week, and we may see the consumer version of the technology by the holiday season. Keep in mind that this, by no means, is a confirmation of a late 2021 launch. It just means that the Ryzen 6000/5000 XT parts will land sometime around the holiday season, either during or after Christmas, possibly even in January.

It’s unclear whether AMD will label Zen 3D as Ryzen 6000 or Ryzen 5000 XT, but one thing is certain. It’ll be compatible with AM4, and we won’t see an entire product stack, just one or two SKUs or three at best. A Ryzen 9 5900XT and 5950XT are my best bets, as the impact of additional L3 cache will be more pronounced with multiple chiplet SKUs.

In gaming workloads, the Ryzen 9 5900X sees a notable gain of up to 25% with V-Cache. In the data center space, different workloads react differently to cache size, clocks, and compute density (cores). Certain applications will see a large gain, while others will not be affected as much. Therefore, it remains to be seen how AMD pitches its Milan-X SKUs, and just how costly they’ll be compared to the existing Milan parts. Power, efficiency, and area are certainly going to be key factors, but other than that we don’t know much.

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Your email address will not be published.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 *

Back to top button