AMD Ryzen 7 5800X3D On Linux: Not For Gaming, But Very Exciting For Other Workloads

Written by Michael Larabel in Processors on 25 April 2022 at 07:28 AM EDT. Page 8 of 8. 56 Comments.

The AMD Ryzen 7 5800X3D didn't prove to be compelling for Linux gaming based on the tests I've carried out thus far. There was just one title (Deus Ex: Mankind Divided) where it was particularly beneficial but for the other mix of Linux native games and Windows titles running by way of Steam Play with Proton+DXVK, there wasn't any significant benefit -- especially to justify the third higher price-tag of the Ryzen 7 5800X3D.

While missing out on AMD's advertised target market for the Ryzen 7 5800X3D, finding significant uplift in other areas made me quite excited from the technical Linux perspective. Particularly for AI/ML workloads tested across Intel oneDNN, Microsoft ONNX, NCNN, Lczero, and other benchmarked workloads showed substantial benefit to the large L3 cache provided by AMD 3D V-Cache even when it meant slightly lower clock speeds. Other HPC benchmarks like ASKAP, OpenFOAM, Incompact3D, and others benefited significantly too if relying on lower-cost Ryzen hardware for budget-constrained compute environments. For general purpose workloads there wasn't much advantage to the Ryzen 7 5800X3D -- one of the main areas seen was with much better Zstd compression performance, which may be useful if doing a lot of compression but otherwise limited in scope.

But for now the Ryzen 7 5800X3D remains AMD's only consumer SKU with 3D V-Cache. For many of the high performance workloads and AI that scale well with core/thread counts will be better off going for a Ryzen 9 5950X with double the core/thread count. When a SKU comes out with 3D V-Cache and matches the top-tier core count on the consumer side, that will be really interesting and could open up such a product to appearing in more technical computing environments for workstations / developer boxes / etc. If cost is not a factor, you could always build a developer workstation right now around Milan-X but it's starting SKU is the EPYC 7373X at $4185 USD.

Given the success of 3D V-Cache in many AI/ML workloads, seeing an AMD Ryzen Embedded offering with 3D V-Cache for edge computing would be interesting. Many of these workloads also will carry over for AMD's increasing focus on Ryzen for dedicated servers.

Anyhow, from some 285 benchmarks carried out under Linux and ignoring the Linux gaming mess, the side-by-side comparison for workloads with a measurable difference included (benchmarks with less than a 2% difference either way omitted):

AMD Ryzen 7 5800X3D Linux Gaming

See all the individual benchmarks in full via this OpenBenchmarking.org result page.

For workloads able to properly utilize AMD #D V-Cache, the 96MB L3 cache on the Ryzen 7 5800X3D really pays off. But 61% of the time, the Ryzen 7 5800X was faster particularly for all the general purpose workloads/benchmarks:

AMD Ryzen 7 5800X3D Linux Gaming

Or the geometric mean of the Ryzen 7 5800X3D is just 3.9% faster than the Ryzen 7 5800X:

AMD Ryzen 7 5800X3D Linux Gaming

Over the span of all the benchmarks conducted, the Ryzen 7 5800X3D did top out at consuming 133 Watts (compared to 105 Watt TDP rating) while the average CPU power consumption across the wid mix of workloads was around 73 Watts.

AMD Ryzen 7 5800X3D Linux Gaming

Linux gamers should forego this CPU given the lack of current games being seemingly unable to benefit from the large L3 cache. But outside of that it does feel that this product is more of a "technical preview" for AMD 3D V-Cache on the consumer/prosumer side. It's a pity there isn't any model akin to a "Ryzen 9 5950X3D" for being able to leverage the highest available core/thread count with the most L3 cache to really shine and fill the void between the $449+ Ryzen 7 5800X3D and the ~$4185 USD EPYC 7373X. The AMD Ryzen 7 5800X3D should be useful to some developers wanting to work on optimizing their codebases for more efficient cache usage and profiling while not shelling out much more money on EPYC Milan-X to use a developer workstation. Or one of the areas where Ryzen 3D V-Cache could be quite interesting is with edge computing / "embedded" given the results I am seeing from the large L3 cache across numerous AI/ML workloads. The many benchmarks I've carried out so far in the few days with the Ryzen 7 5800X3D have shown many interesting areas for potential benefit from a hefty L3 cache in the Ryzen space, which leaves me quite excited for future AMD Ryzen 3D V-Cache products.

AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming
AMD Ryzen 7 5800X3D Linux Gaming

With some preliminary tests carried out on a Ryzen 9 5950X in comparison to the Ryzen 7 5800X3D, in some of the cases really making great use of the AMD #D V-Cache it was enough to slightly put this new CPU past the flagship 5950X. (Stay tuned for more comparison CPU results in future larger article.)

If you are interested in AMD 3D V-Cache for heavy HPC workloads, see my prior AMD EPYC 7773X Linux benchmarks.

Stay tuned for many more AMD Ryzen 7 5800X3D benchmarks over the days/weeks ahead given the limited time I've had so far for working with this CPU since having to purchase it retail last week and exploring more areas where the 3D V-Cache may pay off. If you enjoy all my Linux hardware testing, consider showing your support via joining Phoronix Premium (or PayPal tips), or at least not using any ad-blocker that is particularly important for reviews such as this where needing to purchase the hardware under test.

If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.


Related Articles
About The Author
Michael Larabel

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.