AMD Ryzen 5000 — especially the 5600X — is blowing my mind
November 14, 2020
231         0

by admin

I knew AMD was going to deliver excellent performance with its next-generation Zen 3 CPU architecture, but I wasn’t ready for this. The new Ryzen chips are fast and powerful, both for content creation and productivity. This isn’t a surprise. The Zen 2 Ryzens are also great for those tasks. But gaming is where the Ryzen 5000-series products see the biggest gains — and it turns AMDs processors into the obvious and best choice on desktop.

AMD is launching the 16-core Ryzen 9 5950X, the 12-core Ryzen 9 5900X, the 8-core Ryzen 7 5800X, and the Ryzen 5 5600X with its new architecture. And all of these chips see huge generational leaps in single-core performance. That is what enables these CPUs to pull ahead of Intel, which has long held the top position when it comes to gaming thanks to its fast single-core performance.

I am coming in late to the Zen 3 review period because of a busy season that had me spending a lot of time on Xbox Series X and other major launches. So for this story, I wanted to latch onto gaming. Everyone already knows and expects Ryzen to outperform its Intel Core counterparts in most production tasks. But if you’re gaming 90% of the time on your PC, has Ryzen really pulled even … or even ahead? Well, the single-core Cinebench score suggests that’s the case. All of the new Ryzens outperformed the 10900K by a significant margin.

To determine that, I did two tests. I ran a number of CPU-intensive games at 1080p and medium settings. This will make the CPU the bottleneck and show the difference in performance between AMD and Intel’s offerings. In the other test, I went for something more realistic. Does it matter if you use Ryzen or Intel when you’re gaming at 4K with ultra settings? Or does the GPU bottleneck negate AMD’s performance improvements?

Let’s get to it.

Ryzen 5000 CPUs are usually faster in ideal conditions

At 1080p and lower graphical settings, you can often see the power difference between the new Ryzens and the Intel Core i9-10900K.

In Avengers, every Ryzen either matched or was a few frames faster than the Intel competition. AMD’s CPUs also have significantly better 95% lows, which suggests a smoother experience. As a reminder, 95% lows and 99% lows tell us that the game is running better than this framerate 95% or 99% of the time, respectively.

Avengers is interesting because it has Intel-branded CPU features in its graphic settings, which I turned to max for these tests. And it seems like that didn’t cause any problems for the AMD CPUs.

Teardown, a game about demolishing a voxel-based environment, also saw better performance on most of the AMD chips. The 5600X surpassed an average framerate of 130.

Watch Dogs: Legion was relatively even. I agree with the consensus that this game likes a fast CPU, but the it clearly benefits from clock speeds instead of cores. My 5600X ran it as well as the 16-core 5950X.

Finally, Microsoft’s Flight Simulator does seem to like more cores — until you consider the overclocked 5600X, which was running at 5GHz on all cores. That was nearly as fast as the 5950X.

Ryzen chips even help get more consistent 4K performance

The difference between 1080p and 2160p for benchmarking is that 2160p puts much more of a burden on the GPU. Rendering four times the pixels is the kind of workload that usually causes a traffic jam in the video card, and that means your CPU doesn’t matter as much.

And that was mostly the case across AMD’s Ryzen 5000 product line. In Watch Dogs at 2160p, the 3500X, 3800X, 3900X, and 3950X all essentially tied at 44 frames per second. Avengers, Teardown, and Flight Sim in 4K all saw a similar tie.

The reality is that the differences between the 5600X all the way up to the 5950X don’t matter when the GPU is the bottleneck. And even in games that are notoriously rough on the CPU, the GPU is more important at 4K.

So does that mean that all of this benchmarking is pointless for anyone gaming at 4K? Is Intel effectively the same as AMD at ultra HD resolutions? Well, no — not if you factor in overclocking.

For these tests, I focused on overclocking the 5600X. And with some tweaking, I was able to get significantly better performance from the AMD part. This led to higher framerates in tests even at 2160p.

Even in Flight Sim, which preferred the 10900K at stock speeds, the overclocked 3600X came out ahead of that.

Now, this isn’t about a head-to-head in an overclocking battle. You can get better performance by tweaking the 10900K as well. This was just a quick way of showing that at 4Kyou don’t necessarily need to spend more money than the $300 5600X. It’ll help in some instances, but it’s probably not going to hold you back for games. And considering how easy it was to overclock, it might even improve your 4K gaming.

Teardown showed similar results.

The overclocked 5600X delivered the same benefits in Avengers as well.

Don’t hesitate to get the 5600X for gaming

My big takeaway here is that the new Ryzen CPUs are great — especially the 5600X. This is a slam dunk of a gaming CPU. And if you are gaming on your PC most of the time and then doing your taxes or some occasional web browsing, you don’t need more than this. Spend $300 on this chip and then put the rest of your budget into a GPU.

If you want to game and livestream from one PC, or if you want to edit video or build 3D models, then start going into the 8-core 5800X or the 12-core 5900X. Even the 5950X, which is $800, seems like a deal when you consider it could give you similar results to a two-PC streaming setup with just one machine.

As for Intel, it’s simply no longer in the conversation. It will continue to fight back, but I won’t be recommending their CPUs to anyone anytime soon. It’s not even like the 10900K or 10600K are bad. They’re actually really good gaming CPUs. It’s just that the equivalent Ryzens are better in every way.

subscribe for YouMedia Newsletter

Leave a Reply

Your email address will not be published. Required fields are marked *

subscribe for YouMedia Newsletter