We’re not talking just any cache, AMD’s new 3D V-Cache technology allows them to stack layers of cache on top of each other to increase the total L3 cache considerably. Thus, the 5800X3D has a whopping 96MB of L3 cache as compared to the normal 5800X’s mere 32MB. The 5800X3D does have slightly lower clocks to accommodate for the extra cache and AMD disabled overclocking on the chip, two important distinctions that make it even more impressive when you see what the chip was able to do in a real-world gaming benchmark.
Benchmark details
XanxoGaming, a Peruvian news outlet, got their hands on the Ryzen 7 5800X3D and decided to test it (and publish those results) ahead of embargo, citing that they are “already banned by AMD” so they don’t really care. We first got to see some synthetic benchmarks yesterday that showed the 5800X3D performing surprisingly bad, but as critics pointed out, the chip was never meant for or marketed to be better at anything other than gaming. Today, XanxoGaming followed up on their benchmarks from yesterday with a proper gaming benchmark to give the 5800X3D a chance to flex its muscle. XanxoGaming conducted this test in conjunction with the developer of CapFrameX, a popular monitoring tool used to analyze frames in games. The Ryzen 7 5800X3D was tested in Shadow of the Tomb Raider at 720p using custom settings purposefully tuned to make the game more CPU-intensive and to remove GPU dependency, so a fairer test could be conducted. Up against the 5800X3D was Intel’s Core i9-12900K, the company’s fastest mainstream processor, and the Core i9-12900KS, the pre-binned version of the same CPU that boosts up to 5.5Ghz out of the box. Conducting the test at 720p, CapFrameX metioned “GPUs don’t really matter at this low resolution, so it’s very close to an ‘apple to apple’ comparison. Very close.” Before we get to the benchmark, I should go over the test benches real quick as they are not identical across both platforms. The Intel system is equipped with a Core i9-12900K and 12900KS, with a GeForce RTX 3090 Ti and CL40 DDR5 memory running at 4800Mhz. On the contrary, the AMD system is kitted with an RTX 3080 instead, and CL14 DDR4 memory running at 3200Mhz.
Ryzen 7 5800X3D vs. the competition
Getting to the benchmark itself, the Ryzen 7 5800X3D was able to achieve an average of 231 FPS in Tomb Raider. Intel’s Core i9-12900K, on the other hand, achieved just 190 FPS whereas the i9-12900KS was able to take that number up to 200 FPS. As you can tell, the 5800X3D managed to score at least 31 FPS higher than its Intel contemporaries in the same test. Putting those numbers into a calculator, we find out that the Ryzen 7 5800X3D is 16% faster than the Core i9-12900KS and 20% faster than the Core i9-12900K, at least in this one game. That sits right with AMD’s claims that the 5800X3D is around 10% faster than the Core i9-12900K ate 1080p resolution with high-quality settings. While one test alone shouldn’t be enough to warrant it, the 5800X3D being the “world’s fastest gaming processor” certainly doesn’t seem to be a lie so far. That result is even more impressive when you consider how big an advantage the Intel system had. If the test was even slightly GPU-reliant, the RTX 3090 Ti-equipped Intel system would’ve destroyed the RTX 3080 inside the Ryzen system. Even with a testbed advantage, AMD prevailed at the end and that gives me immense hope that this chip and the 3D V-Cache AMD is developing is truly something special. Not only that, but, as mentioned before, the Ryzen 5800X3D can’t even be overclocked and has much lower clock speeds than the Core i9-12900K and 12900KS. The Ryzen 7 5800X3D can only boost up to 4.5Ghz, and it can’t be manually overclocked to go over that in any way, whereas the 12900KS has a 5.5Ghz boost speed and can actually be overclocked to potentially go even higher. On top of that, the 12900K(S) is a 14-core, 24-thread processor in comparison to just 8 cores and 16 threads on the 5800X3D. Clock speeds along with single-threaded performance are often the most important constituents when it comes to CPU gaming performance, but AMD has proved that extra cache can go a long way in producing more frames, even with inferior clock speeds and core counts.
Further testing is needed
While the results seen above paint the processor in an exceedingly positive light, we simply cannot base our judgement of an unreleased processor on just one test that isn’t even exactly apples-to-apples. For instance, CapFrameX’s own testing of the Core i9-12900K with a much faster 6400Mhz kit of DDR5 memory enabled it to net 220+ FPS in certain games so the use of a slower kit and evasion of a GPU bottleneck are probably limiting the full potential of the Intel offering. Realistically, nobody will be playing at 720p with a Ryzen 7 5800X3D or a Core i9 processor. Therefore, the GPU would have to get involved in some way and that can sway results in either Blue or Red Team’s favor. That being said, this is probably the most fair and just foundation for a benchmark as it eliminates any GPU reliance and puts all the focus on the CPUs itself. We just have to wait and see more tests from different reviewers to truly evaluate if the 5800X3D really is the world’s fastest gaming processor or just a lucky opportunist.
