Rumor: AMD may be planning a 16-core, 32-thread Ryzen titan to challenge Intel’s Broadwell-E, Skylake-X

PT KONTAK PERKASA FUTURES - AMD’s Ryzen 7 1800X, 1700X, and 1700 all debuted earlier this month, officially putting Intel on notice that Team Green was back and ready to fight. Now, there’s a rumor making the rounds that Ryzen 18 may not be the end of AMD’s desktop ambitions. Supposedly, there’s a chip coming that would go toe-to-toe with Intel’s HEDT line.

Obligatory reminder: Take this with enough salt to kill a donkey. Here’s the text of the rumor, as reported by FLCLimax, at Overclock.net:

Public knowledge by now but AMD has a new HEDT platform coming out in a couple of months. You’ll see more of it at Computex I believe.
It’s a 16 core /32 Thread, quad channel behemoth. And it is insanely quick in the tests that Ryzen is already excelling at. So Cinebench, and all other related productivity programs. The gaming issues that were causing the Ryzen AM4 CPUs to behave erratically to say the least have been ironed out. It’s akin to a newer revision on a newer platform. This should be competing with the Xeon and of course 6950X Intel offers for $1700~$1800USD, but at about $1,000 USD if not less for some Skews(sic). Coming soon. CPSs(sic) are pretty big physically, about twice the size of surrent (sic) 6950X CPUs and a bit more perhaps. And if you were hoping for pins, nope it’s strictly LGA! IT’s NOT 8 channel, but Quad.

One thing I want to call out right from the start. FLCLimax’s claim that “the gaming issues… have been ironed out,” is not, to the absolute best of my knowledge, true. It’s true some motherboards had early BIOS problems that made them perform more poorly than they should’ve in gaming benchmarks. AMD said this issue is because games compiled prior to Ryzen do not accurately detect its capabilities or take advantage of the CPUs architecture. And AMD has already said the issue isn’t the Windows 10 scheduler, despite some users being convinced this was the case. Some games may indeed be ironed out if developers release patches to improve Ryzen’s performance, but the issue isn’t in the thread scheduler and it doesn’t seem to be something AMD can improve on its own.

AMD has shown how it connects two Naples chips together with its Infinity Fabric, but connecting 16 cores together in a single socket is still difficult if you want to minimize bottlenecks.

Now, as for the rest of the rumor, it’s by no means impossible. AMD is already planning to build a 32-core Ryzen with 8 memory channels (one per CCX). A 16-core device could easily turn out to be a quad-channel part, and that would compete well against Intel’s HEDT lineup.

The one thing I’d caution against, however, is treating this rumor as if its either proof AMD will launch such a chip or that this kind of processor would be desirable to most enthusiasts. First, AMD will be required to cut clock speeds. The more cores you have, the lower your maximum clock is going to be. That’s always been true for server chips from both companies. Even allowing for some variation between product SKUs, higher core counts mean lower clock speeds or similar clock speeds at a substantially higher TDP. Intel’s E7-8894 v4 (24 cores) has a base clock of 2.4GHz and a boost clock of 3.4GHz in a 165W TDP. The E7-8891 v4 (10 cores) has a base clock of 2.8GHz and a boost of 3.5GHz in the same 165W power envelope.

Which of these CPUs is “better” depends entirely on how multi-threaded your workloads are, of course, but the trend holds clear. If AMD wants to build a 16-core Ryzen, even in a 150W TDP, it’s going to have to give up some clock speed to get there. It might make a fabulous workstation chip, but I’d bet you’d get better gaming performance from a different CPU in AMD’s lineup.

Would a 16-core Ryzen punch holes in Intel’s product family? Possibly. There is some concern that by joining that many CCX’s together, the relatively limited 22GB/s of bandwidth between the CCX’s could prove a scaling bottleneck with that many chips to handle. And it’s absolutely possible, even likely, that gamers would see little benefit from this kind of firepower. Games are typically designed to run on higher clocks with lower threads; few platforms can saturate threads to the point that they count for more than raw clocks.

Still, it’ll be interesting to see if this rumor proves true. Depending on what AMD pulls off, it might be able to regain market share, even at the highest end of the non-server market.

Source : extremetech.com