AMD states that there are no technical reasons why it didn’t launch a dual CCD processor with dual 3D V-cache, it came down to performance and pricing. The Ryzen 9000X3D series debuted with the launch of the Ryzen 7 9800X3D which quickly seized the crown as the new king of gaming processors. It also piqued the interest of PC users looking for a new halo CPU that might incorporate a dual CCD package with a dual 3D V-Cache design, but that didn’t happen. Despite early rumors that the Ryzen 9 9950X3D might get such a design, it was revealed to still have a single 3D V-Cache. However, it is the first X3D part to match clock speeds with its non-X3D counterpart.
While the idea of dual 3D V-cache sounds like a great boon to gamers the reality is that there are other issues to consider such as latency from the Infinity Fabric as well as delays brought on by scheduling tasks between the two CCDs. AMD presently uses a scheduler to direct game-related tasks to one CCD for optimal performance so breaking from that could likely add processing delays. One last detail to note is that reviewers have long said there are diminishing returns for increasing the 3D V-Cache as not all games will take advantage of it. AMD states that while it is technically possible to make a dual CCD processor that has a dual 3D V-cache, the gains simply would not justify the cost of such a processor.
Per HardwareLuxx:
“We asked AMD if there were any technical reasons why we haven’t seen a Ryzen processor with two CCDs and 3D V-Cache on each of the CCDs. The answer was surprising: there are no technical reasons or challenges. Such a processor would simply be too expensive and games would not benefit from a second CCD with 3D V-Cache to the same extent as they do from the step from 32 to 96 MB L3 cache for one CCD.”
It’s been said that a special edition X3D processor featuring dual CCD/dual 3D V-Cache could happen at some point but for right now, it just doesn’t make sense for the cost to the consumer. That doesn’t mean something might not pop up down the road but costs will need to decrease enough to justify AMD moving forward with it. There are, of course, PC users who use their computers for more than just gaming and are interested in what such a CPU configuration could offer.
Discussion (4 replies)
Join Discussion →I'm guessing they prototype tested it, and it didn't provide enough uplift to justify the cost. I'd imagine the latency between CCD's and how the Windows scheduler works just didn't work out.
That's a better answer than just saying "Oh people wanted that?" At least they tested and said it wasn't worth the squeeze. 1k chip for on par performance would have pissed people off.
It may not have had benefits for games, but I sure could have found uses for it in distributed computing.
Oh yea totally but they would also want more PCIE lanes, and a ton more ram. I'm not building a hypervisor today limited to 256 gig of ram.
BUT yea that is still a fair point. Though for a home lab a 16 core box with 256 gig would have been fine.