3D V-Cache Infused Zen 4: AMD Ryzen 9 7950X3D Gaming Review

For a look at AMD’s Zen 4 and Intel’s Raptor Lake CPUs in creator workloads, check out our recent in-depth look.

When AMD released its first 3D V-Cache processor with the Ryzen 7 5800X3D, we were left impressed with the level of gaming performance that could be gained with a doubling of cache, and it became obvious that the design would lead to more cache-rich models from both AMD and Intel down-the-road.

This month, AMD rolled out the first two Zen 4-based models that have 3D V-Cache augmentation: a 16-core Ryzen 9 7950X3D and 12-core Ryzen 9 7900X3D. The third option, an 8-core Ryzen 7 7800X3D, is due soon.

We hoped to have had this review published weeks ago, but issues (our fault) plagued the first round of testing. We fine-tuned testing methods, and retested the same chips a second time. This review emphasized to us that CPU-bound gaming benchmarking is some of the most tedious there is. Comparing one GPU to another in gaming is almost dead simple in comparison. CPU-bound gaming testing is really sensitive, and more than the usual number of runs need to be executed to deliver a high level of confidence in the results.

In this review, we have a mix of new and old game titles, to see how a chip like the 7950X3D can benefit gaming performance overall. As was the case with the 5800X3D, high-resolution gamers should opt for a CPU without 3D V-Cache, because the gaming benefit is almost nonexistent above 1440p. Even 1440p can deliver underwhelming results at times – some of which can be seen later.

Before going further, here’s a quick look at AMD’s current-gen Zen 4-based processor lineup:

Despite the fact that AMD’s doubled the available cache in the top-end 3D V-Cache SKUs, continued optimization of the architecture since the original Zen 4 launch has led to these newer chips requiring much less power. It also helps that the 3D V-Cache chips tend to be clocked a bit lower, but that’s not enough to see 50W being shaved off of the TDP. Optimizations like these are always great to see.

Because of the 3D V-Cache infusion, the 7950X3D and 7900X3D enjoy a doubling of L3 cache, from 64MB to 128MB, while the 7800X3D triples the L3 cache over the 7700X, from 32MB to 96MB.

For AMD, it’s easier to augment just one CCD (CPU die; 1 for 8-core and under chips, 2 for higher than 8-core) with lots of cache than both, so the 7900X3D and 7950X3D put all of that extra cache alongside one CCD. The 7800X3D, meanwhile, is just one CCD, so implementation and OS-awareness is a little more straight-forward.

AMD chip shot of its dual-CCD 3D V-Cache processor

Because of the exotic design of the 7900X3D and 7950X3D, Windows needs updated bits in order to ensure that gaming is taking place on the CCD with the extra cache. To help make sure that happens, AMD’s newest chipset software installs an applet which directs gaming threads correctly. If you don’t have the chipset software (or this applet) running, you won’t be able to guarantee that your game will be running on the correct CCD.

AMD’s 3D V-Cache optimizer services installed with chipset driver

In time, those who own either a 7900X3D or 7950X3D will need to do nothing other than update their EFI, Windows, and the AMD chipset driver. For now, users should additionally run the following code in a command prompt as Administrator to force the completion of background tasks.

start "" /wait Rundll32.exe advapi32.dll,ProcessIdleTasks

This could take tens of minutes, and ensures Game Bar is fully updated, as it too includes required optimizations for AMD’s 3D V-Cache chips.

It might also be worth mentioning that these new 3D V-Cache SKUs include integrated graphics just like the rest of the Zen 4 lineup. Considering the target audience of X3D chips, an IGP is going to be largely ignored, but it’s still nice to have one available just in case you’re suddenly left without a discrete GPU for whatever reason.

Something we do find a bit odd with this Radeon IGP is that Windows will try to automatically install a driver for it, even if the user doesn’t intend to use it. We’ve not experienced that kind of behavior with Intel’s CPUs, which have required explicit installation of the IGP driver. On AMD, this behavior complicated the LuxMark test in our recent creator performance deep-dive, as it was trying to use the IGP instead of the discrete GPU. Our fix was to simply uninstall the IGP driver Windows put there, and then configure the OS to stop automatically installing drivers.

Alright… let’s take a look at the test rigs, our methodologies, and Cinebench scores ahead of each CPU’s game test run. Uninterested? Hit up the next page instead.

Techgage’s CPU Testing Platforms

Testing Considerations

We use an up-to-date Windows for our testing, as well as the latest chipset driver for each respective platform. In the pursuit of accurate, repeatable benchmarks, here are some basic guidelines we follow:

• Disruptive services are disabled; eg: Search, Cortana, User Account Control, Defender, etc.
• Overlays and / or other extras are not installed with the graphics driver.
• Vsync is disabled at the driver level (and in any tested game or application).
• Default (usually Balanced) power profile is used (with screen and sleep timeouts disabled).
• OSes are never transplanted from one machine to another.
• We validate system configurations before kicking off any test run.
• Testing doesn’t begin until the PC is idle (keeps a steady minimum wattage).
• All tests are repeated until there is a high degree of confidence in the results.
• Only one game title is tested per OS boot to prevent phantom processes from interfering.
• We do not use outdated data. Performance numbers seen in this article are current.

Before each one of our tested processors went through the game test gauntlet, we ran Cinebench to gauge their overall potential:

Based on these rendering results, Intel’s Core i9-13900K looks mighty – but it is mighty enough to compensate for AMD’s cache buff? We’ll begin our investigation on the next page.