The AMD Ryzen 7 5800X3D Review: 96 MB of L3 3D V-Cache Designed For Gamers
by Gavin Bonshor on June 30, 2022 8:00 AM EST- Posted in
- CPUs
- AMD
- DDR4
- AM4
- Ryzen
- V-Cache
- Ryzen 7 5800X3D
- Zen3
- 3D V-Cache
CPU Benchmark Performance: Simulation And Rendering
Simulation and Science have a lot of overlap in the benchmarking world, however for this distinction we’re separating into two segments mostly based on the utility of the resulting data. The benchmarks that fall under Science have a distinct use for the data they output – in our Simulation section, these act more like synthetics but at some level are still trying to simulate a given environment.
We are using DDR4 memory at the following settings:
- DDR4-3200
Simulation
Rendering
As we mentioned on the last page, where the Ryzen 7 5800X3D can use its 96 MB of L3 V-Cache to improve performance, it does, such as in our Factorio test. For the most part, it doesn't do enough in most situations to improve performance when compared to the Ryzen 7 5800X.
CPU Benchmark Performance: Power, Office, And Science
CPU Benchmark Performance: Encoding and Compression
Facebook
Twitter
RSS
125 Comments
View All Comments
asmian - Friday, July 1, 2022 - link
No, I have to agree, this was constantly confusing for me. Most stand-out colour MUST be the thing that you're reviewing. Having a near comparison be the brightest or most eye-catching is very misleading.hfm - Friday, July 1, 2022 - link
I agree as well. There was also a ton of typos and errors. They need a second pair of eyes on these things.MDD1963 - Thursday, June 30, 2022 - link
What a prompt, timely review! :/TeXWiller - Thursday, June 30, 2022 - link
For some impressive engineering software results take a look at the Phoronix's tests. One thing I personally miss in most reviews is not using at least the "slower" preset for the x265 and trying the deblocking filters, particularly the EEDI2 for a low core-count performance test. Using the faster presets quickly turn the computing bounded test into a memory bandwidth, or even to a IO bandwidth test.TeXWiller - Saturday, July 2, 2022 - link
I noticed I wrote deblocking when I actually meant deinterlacing there.garblah - Friday, July 1, 2022 - link
The biggest performance gains with the 5800x3d come from games with less than stellar multi core utilization and yet are very demanding from a single core performance standpointMSFS 2020 and DCS, for example, especially in VR, since you have double the number of frames for the CPU to set up for each game frame. The 5800x3d performs 40 percent better than a 5600x in those situations. Dual Rank memory at 3600mhz is important to get the most out of it, too.
Escape from Tarkov is another one with huge gains from the 5800x3d.
Oxford Guy - Saturday, July 2, 2022 - link
‘ As a result, in lieu of CPU overclocking, the biggest thing a user can do to influence higher performance with the Ryzen 7 5800X3D is to use faster DDR4 memory with lower latencies, such as a good DDR4-3600 kit. These settings are also the known sweet spot for AMD's Infinity Fabric Interconnect as set out by AMD.’Perhaps it’s a bit odd, then, to see DDR-3200 listed below that paragraph — the apparently only RAM chosen for the testing.
mode_13h - Saturday, July 2, 2022 - link
> 5800X3D Compute Analysis: Extra L3 Does Little For Compute PerformanceOr, maybe your testing of "compute performance" is too limited to find the cases where it's a big win.
Phoronix found many cases where the extra L3 cache is a substantial win for technical computing.
https://www.phoronix.com/scan.php?page=article&...
What we see is that there are plenty of cases that benefit from it, substantially. As mentioned in this article, they found cases where the extra L3 cache is enough even for it to pull ahead of the i9-12900K, even when it beats all of AMD's other desktop CPU models.
MDD1963 - Monday, July 4, 2022 - link
Let's hear it for timely reviews!!!mode_13h - Tuesday, July 5, 2022 - link
They know it's late, but I'd rather have it than not. Even though there are some gaps in their testing, and some of the usual analysis is lacking (e.g. latency analysis), it does give us some apples-to-apples data vs. other CPUs they have or will review.