For our first review of the Ryzen 5000, we looked at the Ryzen 9 5950X, AMD's new AM4 processor with 16 cores and 32 threads and $ 800. Today we're looking at the Ryzen 9 5900X, the 12-core, 24-thread CPU that retails for $ 550.
That makes the 5900X ~ 8% cheaper per core compared to its bigger brother, as AMD can get away with a premium for the 16-core model as it doesn't face any real competition.
The 12-core R9 5900X can keep up with the Intel Core i9-10900K. Typically, the 10-core Intel processor sells for $ 530, but due to shipping issues, it usually sells for $ 550 or $ 600, which means the 10900K and 5900X are similar in price.
Now we need to figure out how they stack up in terms of performance.
The 5950X is a beast, so we expect good things from the 5900X, the only major change being the lowering of the core count. The clock rates are largely the same, the boost frequency has been reduced by 100 MHz, while the base frequency has been increased by 300 MHz. Since both CCDs are still active, a total L3 cache of 64 MB is packed.
Again, we're not going to look at improvements to the Zen 3 architecture in this test, but rather skip that and jump into the benchmarks.
Before we do that, let's take a quick look at another new AM4 motherboard. In yesterday's review we featured the Asus ROG Crosshair VIII Dark Hero and now we're getting our first look at the MSI B550 Unify, a new B550 board that … well, pretty much contains everything.
Highlights include a 16-phase VRM with Infineon TDA21490 90A power levels, a 6-layer circuit board with 2 ounce thickened copper, elongated aluminum heat sinks with a dedicated VRM heat pipe, four M.2 slots with double-sided M.2 Shield Frozr Cooling and DDR4 A. -XMP support for up to DDR4-5800.
There is also a Unify-X version with only two memory DIMMs. MSI claims that Unify-X can run DDR4-5800 with 4000 G-series processors and DDR4-5300 with 3rd generation Ryzen. While the standard Unify that we have can achieve DDR4-5600 with 4000 G series, they still have to make any claims of Ryzen 5000 series.
Later this month we'll be putting together a detailed guide to Zen 3 storage performance and using the standard Unify for the majority of our testing. This MSI board costs $ 290, making it a Gigabyte B550 Master competitor. After our numerous Ryzen reviews, we will be doing more motherboard tests again.
To make all of our new Ryzen 5000 reviews easy to work with, we're using the same god-like MSI X570 motherboard along with four 8GB G.Skill TridentZ DDR4-3200 CL14 memory modules for a capacity of 32GB. All systems are cooled by the Corsair iCUE H150i Elite Capellix AIO.
We use an RTX 2080 Ti for all productivity tests, but for the gaming benchmarks we updated all of our numbers with the RTX 3090. So let's get to the graphics …
First we have Cinebench R20 and wow, 8487 points is pretty insane. This is a 17% increase over the 3900X, making the 5900X only 8% slower than the 3950X. These are impressive numbers when you consider that the CPU packs contain 25% fewer cores. Compared to the Core i9-10900K, which is about the same price, the new 5900X is 32% faster, even though it has 20% more cores.
Running the single-core benchmark shows why the 5900X is so much faster than the 10900K in Cinebench R20. Here we see a 16% increase in single-thread performance. It's also a 23% improvement over the 3900X, and with a single core score similar to the 5950X, we expect strong gaming performance.
We also watched how the 5900X clocked in each of the Cinebench R20 tests. For the multi-core test, in which all cores are heavily loaded, the 5900X clocked at around 4.3 GHz. In the meantime, AMD has announced a base clock frequency of 3.7 GHz.
In the Cinebench single core test, the 5900X normally ran at 4.9 GHz, which is 100 MHz above the stated specification. AMD has made sure in this case that the maximum boost frequency is not only easily reached, but can also be exceeded in order to avoid the drama with Zen 2 and its boost clocks.
7-zip looks at the compression performance and here the 5900X is 17% faster than its predecessor. It's also able to displace the 3950X, making it 37% faster than the 10900K.
While decompressing, we find that SMT is more useful and therefore the 3950X is able to move forward. However, the 5900X is 8% slower and still offers a significant 18% increase in performance over the 3900X. As a result, it's much faster than Intel's best, beating the 10900K by a whopping 43%.
AES encryption performance was slightly improved with Zen 3, and this is where Zen 2 already excels. So it's not surprising that the 5900X dominates the 10900K by a huge margin of 52%.
At Blender, we're seeing a 12% increase in performance for the 5900X over the 3900X, meaning it was 22% faster than the 10900K.
Compared to the 3950X, it was 13% slower and then 18% slower than the 5950X.
When testing the 5950X, we found that the new Zen 3 architecture is absolutely falling apart in V-Ray, and that's exactly what we're seeing with the 12-core model. Here the 5900X essentially matched the older 16-core 3950X, and that meant it was 23% faster than its predecessor and 32% faster than its main competitor, the 10900K, so a bit of a shellacking.
The last rendering benchmark we have is Corona, where the 5900X is much faster than the 3900X, increasing performance by 22%, and delivering 3950X-like performance. This is a huge improvement from gene to gene and also meant the 5900X was 24% faster than the 10900K.
For those of you interested in code compilation performance, the 5900X is only 9% faster than the 3900X. This is one of the smaller wins we've seen so far. In fact, it may be the smallest.
Even so, it managed to beat the 10900K with a convincing margin of 26%.
DaVinci Resolve Studio 16 with the Puget Systems Benchmark actually offers the weakest gains we've seen yet. The 5900X was only 8% faster than the 3900X and roughly the same as the 3950X.
Adobe Premiere Pro 2020 cuts the bar even further. Here the 5900X was only 5% faster than the 3900X while it was only 9% faster than the 10900K. A performance benefit of 9% for roughly the same money is hardly a cause for sneezing, even if it is a worst-case scenario.
The strong single-core performance comes in handy when testing with Photoshop. Here the 5900X matched the 5950X, which meant it was 10% faster than the 10900K and a whopping 20% faster than its predecessor, the 3900X.
In After Effects, another application that is heavily dependent on the performance of individual threads, we see strong performance gains over Zen 2. Here the 5900X was 15% faster than the 3900X and 13% faster than the 10900K and was almost the same as the 5950X.
To measure power consumption, we monitored the entire system consumption in Blender. The Ryzen 9 5900X was the same as the 3900X's power consumption, although both use the same 7nm process, the Zen 3 can boost performance in Blender by 12%. In addition, the 5900X reduced overall system consumption by 23% compared to the 10900K.
In terms of temperatures, the Ryzen 9 3900X and 5900X were both tested in a 21C room on an open test bench with the Corsair iCUE H150i Elite Capellix AIO on the MSI X570 Godlike installed in the Corsair Obsidian 500D.
We could see that the 5900X ran 7 to 8 degrees cooler than the 3900X under the same test conditions, even though it ran the same 7 nm process at a slightly higher clock speed. Part of that will be down to 7nm runtime, but AMD has also included a number of architecture optimizations. These improvements also allow for a slightly lower Vcore, which would have helped reduce power consumption and therefore thermal.
We have the same set of gaming tests as yesterday's 5950X test with a GeForce RTX 3090. In the same order, we start with Far Cry New Dawn.
The 5900X matched the 12850X at 59 fps, making it 6% slower than the 10900K, but still a good overall result given the double-digit percentage improvement over the 3900X.
The 5900X follows the 10900K in Rainbow Six Siege by a small 5% lead. We see the CPU match the 5950X for an 8% increase over the 3900X. This was enough to supplant some of the lower core count Intel processors like the 10700K.
In Watch Dogs: Legion, we see strong performance from the 5900X, matching the 10700K and 10900K with just 2 fps. This is also a pretty big 15% increase over the 3900X, which even battled the Core i5-10400 at this low resolution.
We're also seeing big performance gains in the 2020 F1, as the 5900X is 21% faster than the 3900X and reaches 272 fps with the RTX 3090. That's a great result with the new 12-core Ryzen processor just 2% behind the 10900K. That's close enough to believe a tie in our book.
The performance increase in Horizon Zero Dawn is solid. Here the 5900X is 15% faster than the 3900X and can decisively keep up with the 10900K in that it has the same performance for a few frames.
The performance in Borderlands 3 is only slightly improved. The 5900X was only 5% faster than the 3900X, although that was enough to put it within range of the 10900K and only lose 4%.
As mentioned in the 5950X review, these Zen 3 processors are an absolute must for Death Stranding. The 5900X matched the 16-core model exactly, making it 12% faster than the 10900K and an incredible 35% faster than the 3900X.
The 5900X can also compete with the 5950X in Shadow of the Tomb Raider, which means it's 4% faster than the 10900K. A small win close enough to make a tie. The new processor beats the old 3900X by 26% in this title.
Hitman 2's performance is also solid. Here the 5900X couldn't compete with the 5950X, but was still 7% faster than the 10900K and 34% faster than the 3900X.
Performance in Star Wars: Squadrons is strong and on par with the 10900K, although it is convincing for 1% low performance. Then we see a 27% increase in performance over the 3900X, with the average frame rate increasing from 247 fps to 314 fps.
The results of Serious Sam 4 look good for Ryzen. The 5900X is 11% faster than the 10900K and an amazing 47% faster than the 3900X.
Average game performance
Here's a look at the average performance of the 11 games we tested. The 5900X is largely the same as the 5950X in gaming, so it's not surprising that both deliver the same average of 214 fps.
This also means that the 5900X and 10900K are overall on par in terms of gaming performance, at least in our small selection of games. With that in mind, we're looking forward to a head-to-head benchmark of 40 games with these processors.
It's also impressive to see a 20% increase in performance over the 3900X. We know AMD has claimed a 19% IPC improvement, but seeing that translate into real-world gaming wins is something else.
With the 5900X, we were able to achieve an all-core overclocking of 4.6 GHz with 1.375 V. This isn't quite as good as the 5950X's 4.7 GHz, but it's a solid all-core overclock nonetheless. This increased the multi-core performance of the Cinebench R20 by 7%. So it's not amazing and nothing more than the 5950X's 20% increase, but it's a free feat provided you have a decent cooler and motherboard.
However, we cannot recommend this overclocking as it will degrade single core performance. The 5900X used to reach 4.9 GHz in this benchmark, now we're limiting it to 4.6 GHz and that reduced the score by 7%.
Overclocking is useful for running core-heavy applications when you want to sacrifice efficiency for more performance. The gains are not surprising, as can be seen in the Cinebench R20 multi-core test. Here, too, we only see a performance increase of 6%.
For this 6% increase in performance, the overall system consumption is increased by 16%, and while this is still 11% less than a standard 10900K, it translates into a reduction in performance per watt.
For the most part, gamers don't want to overclock the 5900X, at least not the way we did. Maybe overclocking with CCX works better. Here we are seeing a 5% reduction in performance, although obviously you will only see this with high CPU tying, which is not often the case in games.
We see a slight performance improvement in Far Cry New Dawn as overclocking increased the average frame rate by 2%.
Time for the price-performance analysis. Starting with Cinebench, we see the 5900X cost 25% less than the 10900K, with the Intel processor priced at $ 550, although we've seen it sell up to $ 600. The price should be around $ 530, but if availability is low, you'll pay a premium to get it.
With a less affordable program like Adobe Premiere, the 5900X still delivers more value than the 10900K because it was 8% less expensive. Not much of a difference, but like I said, this was one of the worst uses for the 5900X.
Although the 5900X currently costs almost $ 100 more than the 3900X, it still costs 5% less per frame. Compared to the 10900K, it was only 4% cheaper per frame.
The average of 11 games looks a little different. Basically, the 5900X offers the same value for CPU-limited gaming as the 3900X, although it costs $ 90 more in our comparison. Compared to the 10900K, we see the same value for games.
What we learned
We have to say, as impressive as the Ryzen 9 5950X is, we found the 5900X's review a little more interesting, largely because it has a direct competitor in the same price range. The Ryzen 9 5900X is a more wallet-friendly CPU, and possibly the most popular of the two for those looking to build a high-end system.
This makes a lot more sense for gamers. The question now is, for gamers looking to build a high-end system, which CPU should they buy, the 10900K or the 5900X? For the vast majority of people, this won't make any real difference. The performance will be about the same, and even if the CPU is limited 99% of the time, you won't be able to tell the difference. Looking ahead, the 5900X is the more powerful of the two models and also has more cores for booting.
If we look at the newer, more sophisticated games like Death Stranding and Serious Sam 4, we see that Ryzen is leading by a good margin. Hence, we foresee a situation in 1 to 2 years' time where the 5900X will lead the 10900K by a comfortable lead in the most demanding titles, although this is our speculation based on what we saw today.
For the most part, gaming will likely limit the GPU, and that means you'll notice the difference between the 10900K and 5900X even less in the short term. The 5900X has the advantage of being more energy efficient. There is also PCI Express 4.0 support in newer motherboards that will come in handy at a later date.
Then the choice is more obvious for those of you running some type of application, especially a core heavy workload. In the worst case, the 5900X is only slightly faster. At best, it's around 40% faster and we've usually seen gains in the 20-30% range.
Right now, Intel seems out of the game in the high-end space, and we suspect things will only get worse for the blue team as we work through the Zen 3 product stack. The Ryzen 9 5900X is a mighty good CPU, and for the 10900K to be considered an option, we don't think it should cost more than $ 500 or less, given the lack of PCIe 4.0.
Stay tuned tomorrow as we review the Ryzen 7 5800X.