AMD today announced its new Ryzen 7000X3D high-end desktop processors to much fanfare, with availability slated for February 2023, you can read all about them in our older article. In our coverage, we noticed something odd about the cache sizes of the 12-core 7900X3D and 16-core 7950X3D. Whereas the 8-core, single-CCD 7800X3D comes with 104 MB of total cache (L2+L3), which works out to 1 MB L2 cache per core and 96 MB of L3 cache (32 MB on-die + 64 MB stacked 3DV cache); the dual-CCD 7900X3D and 7950X3D was shown with total caches of 140 MB and 144 MB, while they should have been 204 MB or 208 MB, respectively.

In our older article, we explored two possibilities—one that the 3DV cache is available on both CCDs but halved in size for whatever reason; and the second more outlandish possibility that only one of the two CCDs has stacked 3DV cache, while the other is a normal planar CCD with just the on-die 32 MB L3 cache. As it turns out, the latter theory is right! AMD put out high-resolution renders of the dual-CCD 7000X3D processors, where only one of the two CCDs is shown having the L3D (L3 cache die) stacked on top. Even real-world pictures of the older "Zen 3" 3DV cache CCDs from the 5800X3D or EPYC "Milan-X" processors show CCDs with 3DV caches having a distinct appearance with dividing lines between the L3D and the structural substrates over the regions of the CCD that have the CPU cores. In these renders, we see these lines drawn on only one of the two CCDs.

AMD today announced its Ryzen 7000X3D "Zen 4" desktop processors with 3D Vertical Cache technology. With these, the company is claiming to have the world's fastest processors for gaming. The company claims to have beaten the Intel Core i9-13900K "Raptor Lake" in gaming, by a margin it feels comfortable to remain competitive with against even the upcoming Core i9-13900KS. At the heart of these processors is the new "Zen 4" 3D Vertical Cache (3DV cache) CCD, which features 64 MB of L3 cache stacked on top of the region of the "Zen 4" CCD that has the on-die 32 MB L3 cache. The 3DV cache runs at the same speed as the on-die L3 cache, and is contiguous with it. The CPU cores see 96 MB of transparent addressable L3 cache.

3DV cache is proven to have a profound impact on gaming performance with the Ryzen 7 5800X3D "Zen 3" processor that helped it beat "Alder Lake" in gaming workloads despite "Zen 3" being a generationally older microarchitecture; and AMD claims to have repeated this magic with the 7000X3D "Zen 4" series, enabling it to beat Intel "Raptor Lake." Unlike with the 5800X3D, AMD don't intend to make gaming performance a trade-off for multi-threaded creator performance, and so it is introducing even 12-core and 16-core SKUs, so you get gaming performance alongside plenty of muscle for creator workloads.

Microsoft and AMD on Thursday released software updates that fix the two performance issues affecting AMD Ryzen processors with Windows 11. The two issues were abnormally high L3 cache latency, and a broken "Preferred Cores" system. The companies had assessed that the issues impact performance of Ryzen processors on Windows 11 by as much as 15%.

The two issues are fixed in separate methods. The L3 cache latency bug is improved through a Windows Update patch, which has been released now as an Update Preview (an Update Preview is not a "beta," but a software update released ahead of its designated "patch Tuesday"). The Update Preview is chronicled under KB5006746, and Windows 11 systems updated with this, get their OS build version set as "build 22000.282." The next update restores the Preferred Cores mechanism that leverages UEFI-CPPC2. This update comes in the form of an AMD Chipset Software update. You'll need to download and install both of the following:

DOWNLOAD: Windows 11 October 21, 2021 Update Preview KB5006746 | AMD Chipset Driver Software 3.10.08.506

Microsoft on October 12 put out the first Cumulative Updates for the new Windows 11 operating system, since its October 5 release. The company's monthly update packages for Windows are unofficially dubbed "patch Tuesday" updates, as they're scheduled to come out on the second Tuesday of each month. Shortly after Windows 11 launch, AMD and Microsoft jointly discovered that Windows 11 is poorly optimized for AMD Ryzen processors, which see significantly increased L3 cache latency, and the UEFI-CPPC2 (preferred cores mechanism) rendered not working. In our own testing, a Ryzen 7 2700X "Pinnacle Ridge" processor, which typically posts an L3 cache latency of 10 ns, was tested to show a latency of 17 ns. This was made much worse with the October 12 "patch Tuesday" update, driving up the latency to 31.9 ns.

AMD put out a statement on social media, which surfaced on Reddit. The company stated that patches for the two issues have been developed, and specified dates on which they'll be released. The patch for the Preferred Cores (UEFI-CPPC2) bug will be released on October 21. Customers can request the patch even earlier. By "customers," AMD is probably referring to big enterprise customers running mission-critical applications on Threadripper or EPYC-powered workstations. The L3 cache latency bug will be fixed through the Windows Update channel, its release is scheduled for October 19.

Apparently, AMD processors officially compatible with Windows 11, exhibit a three-times increase in L3 cache latency with the new operating system. The new operating system is also found to break the "preferred cores" system on AMD processors (UEFI CPPC2), in which the two "best" CPU cores, which can sustain the highest boost frequencies, are highlighted to the operating system, so most of the light-threaded traffic could be sent to them.

AMD and Microsoft jointly made this discovery, and listed out potential impact on application performance. The increased L3 cache latency affects performance of applications sensitive to memory performance. They also warn of a 10-15% loss in gaming performance. On the other hand, a dysfunctional "preferred cores" system would mean reduced performance in light-threaded tasks as the OS is unaware which are the processor's two best cores. Thankfully, both issues can be fixed via software updates, and AMD is working with Microsoft to push fixes for both issues through Windows Update, in an update rollout scheduled within October 2021.

AMD in a blog post earlier today explained that there is no 1:1 correlation between the "best core" grading system displayed in Ryzen Master, and the "preferred cores" addressed by the Windows 10 Scheduler using CPPC2 (Collaborative Power and Performance Control 2). Deployed through BIOS and AMD chipset drivers, CPPC2 forms a middleware between OS and processor, communicating the system's performance demands at a high frequency of 1 ms (Microsoft's default speed for reporting performance states to processors is 15 ms). Ryzen Master, on the other hand, has had the ability to reveal the "best" cores in a Ryzen processor by ranking them across the package, on a CCD (die), and within a CCX. The best core in a CCX is typically marked with a "star" symbol on the software's UI. The fastest core on the package gets a gold star. Dots denote second fastest cores in a CCX.

Over the past couple of months we've posted several investigative reports by our Ryzen memory overclocking guru Yuri "1usmus" Bubly, and a recurring theme with our articles has been to highlight the discrepancy between the highest performing cores as tested by us not corresponding to those highlighted in Ryzen Master. Our definition of "highest performing cores" has been one that's able to reach and sustain the highest boost states, and has the best electrical properties. AMD elaborates that the CPPC2 works independently from the SMU API Ryzen Master uses, and the best cores mapped by Ryzen Master shouldn't correspond with preferred cores reported by CPPC2 to the OS scheduler, so it could send more workload to these cores, benefiting from their higher boosting headroom.