Date: February 4, 2020
Author(s): Rob Williams
AMD’s latest gaming GPU to join its Radeon Navi family is the RX 5600 XT. With its price-point, and 6GB framebuffer, the RX 5600 XT takes NVIDIA’s GeForce RTX 2060 head-on. We’re taking a look at the card with both its reference BIOS, as well as its performance-boosting BIOS, which will bump both the GPU and memory clocks.
AMD had quite a bit to talk about at this year’s CES, but not surprisingly, the lion’s share of the attention went towards the company’s upcoming 64-core Ryzen Threadripper 3990X. The company came prepared with some graphics treats, though, having unveiled its latest Navi lineup addition: Radeon RX 5600 XT. Where’s 5800 XT or 5900 XT? “Big Navi” is due this year (maybe), so it seems likely those will be intertwined.
At the launch of the 5600 XT, there was quite a bit of confusion about what it was supposed to be. Ahead of the embargo lift, AMD asked reviewers to update the BIOS on their provided sample, which in the case of our Sapphire Pulse resulted in a ~10% performance uplift.
RX 5600 XT reference specs are 1375MHz Game Clock, 1560MHz Boost Clock, and 12Gbps memory. That configuration hits a total graphics power (TGP) of 150W. The BIOS we were provided boosted that TGP to 160W, a result from improving the memory to 14Gbps, and increasing the Game Clock to 1615MHz, and Boost Clock to 1750MHz.
At launch, it seemed fair to dump the results generated from the old BIOS, but in reality, those are still “reference” for this GPU. Because of that, we’re including performance from both the BIOS with reference clocks, and also the upgraded version, which can be grabbed from Sapphire’s website. As of the time of writing, all of the Pulse models we find at etail reference the 12Gbps clocks, so you will assuredly have to manually flash the BIOS after purchase (which is not difficult).
Here’s AMD’s current lineup:
|AMD’s Radeon Gaming GPU Lineup|
|Cores||Base MHz||Peak FP32||Memory||Bandwidth||TDP||Price|
|RX 5700 XT Anniv.||2560||1680||10.1 TFLOPS||8 GB 1||448 GB/s||235W||$449|
|RX 5700 XT||2560||1605||9.75 TFLOPS||8 GB 1||448 GB/s||225W||$399|
|RX 5700||2304||1465||7.95 TFLOPS||8 GB 1||448 GB/s||180W||$349|
|RX 5600 XT||2304||1375||7.19 TFLOPS||6 GB 1||288 GB/s||150W||$289|
|RX 5600||2048||1375||6.39 TFLOPS||6 GB 1||288 GB/s||150W||OEM|
|RX 5500 XT||1408||1717||5.2 TFLOPS||8 GB 1||224 GB/s||130W||$199|
|RX 590||2304||1576||7.1 TFLOPS||8 GB 3||256 GB/s||225 W||$199|
|Notes||1 GDDR6; 2 GDDR5X; 3 GDDR5; 4 HBM2|
Architecture: RX 5500~5700 = Navi; Radeon RX 590 = Polaris
If there’s one strange spec with the RX 5600 XT, it’s the fact that it bundles in 6GB of GDDR6, whereas the lower-end RX 5500 XT sports up to 8GB (although it typically sells with only 4GB). The downgrade in memory comes with the trade-off of offering better bandwidth, from 224GB/s on the 5500 XT to 288GB/s on the reference 5600 XT. One of the 14Gbps 5600 XT models will further boost that to 336GB/s.
With that all covered, let’s move onto a quick look at our test setup, and then get right into testing:
|Techgage Gaming GPU Test PC|
|Processor||Intel Core i9-9900KS (4.0GHz Base, 5.0GHz Turbo, 8C/16T)|
|Motherboard||ASUS ROG STRIX Z390-E GAMING|
CPU tested with BIOS 1401 (November 26, 2019)
|Memory||G.SKILL TridentZ Royal (F4-3600C16-8GTRG) 8GB x 2|
Operates at DDR4-3600 16-16-16 (1.35V)
|AMD Graphics||AMD Radeon RX 5700 (8GB; Adrenaline 20.1.2)|
AMD Radeon RX 5600 XT (6GB; Adrenaline Review Driver)
AMD Radeon RX 5500 XT (8GB; Adrenaline 20.1.2)
AMD Radeon RX 590 (8GB; Adrenaline 20.1.2)
|NVIDIA Graphics||NVIDIA GeForce RTX 2060 (6GB; GeForce 441.87)|
NVIDIA GeForce GTX 1660 Ti (6GB; GeForce 441.87)
NVIDIA GeForce GTX 1660 SUPER (6GB; GeForce 441.87)
NVIDIA GeForce GTX 1660 (6GB; GeForce 441.87)
|Storage||GeForce: WD Blue 3D NAND 1TB (SATA 6Gbps)|
Radeon: Kingston SSDNow V310 960GB (SATA 6Gbps)
|Power Supply||Corsair RM650x (650W)|
|Chassis||NZXT S340 Elite Mid-tower|
|Cooling||Corsair Hydro H100i V2 AIO Liquid Cooler (240mm)|
|Et cetera||Windows 10 Pro (64-bit; 1909)|
All of the GPUs have been tested with current (as of the time the round of testing began) drivers, and with an up-to-date Windows 10 (1909). Our operating system is kept clean and optimized to reduce benchmark interference, ensuring accurate results. V-Sync, G-SYNC, and FreeSync are disabled at the monitor and driver level. Both Intel’s chipset driver and Management Engine (ME) are updated to the latest versions.
A total of eight games are included in this particular test suite. Here’s the full list, along with their developer allegiances:
Note: You can download all of the tested setting images at once here (ZIP, 4MB).
With our first result, we can see a noticeable gain between the RX 5600 XT’s reference clocks, and those that Sapphire has upgraded. This is actually one of the tamer performance deltas between the configurations, but it’s about what we’d expect to see in general between a reference and a pre-OCed card.
Sadly, ever since Borderlands 3 came out, we’ve heard multiple reports of the game not running ideally over the long haul on either AMD or NVIDIA, but we’ve never been able to nail down where it’s worse. We’ve been waiting for the game to hit Steam before playing it, so we’re hoping by that point, some of these long-standing bugs will finally be nailed down.
If the 5600 XT delivers 82 FPS at an SRP of $289, then the RX 5700 at $349 begins to look a little questionable. That’s especially the case if we’re taking a look at the boosted clocks on the Pulse. Notably, in this particular game, AMD’s $289 beats out NVIDIA’s $299 RTX 2060 – a card that was $349 ahead of AMD’s 5600 XT launch.
With F1 2019, we’re again seeing the RTX 2060 struggle to exceed the Pulse 5600 XT’s performance, which is starting to make the price drop on the RTX 2060 make a bit of sense. But here, we see a bigger performance delta between the reference-clocked 5600 XT, and this boosted clocked one.
We’re seeing a full 10 FPS difference between the two 5600 XT configurations here, which might seem insignificant at 100 FPS, but it’s really not. When we look at the minimums, that kind of difference makes an even bigger real-world impression. It could be that you want to run this game at 1440p, which means the FPS is going to drop, so any extra performance you can eke out of it will be put to good use.
In both of these charts so far, we’ve seen AMD’s newer RX 5500 XT fall behind its older RX 590 sibling, which makes the newer release seem quite strange in some ways. Granted, Navi is the technically better architecture overall, but Polaris is really battle-tested, so we’d be quicker to recommend that one over the Navi card if it came down to it – and if the price was where it should be. It makes no sense to not go for whichever card makes the most sense value-wise.
The going gets real tough with Metro Exodus. For this test, we use the in-game benchmark, which suffers more dramatic lows than we see in most of our other testing. We don’t really think that this benchmark is representative of the entire game, but it seems to be good at testing its worse-case scenario, especially with the minimums.
That said, this is the third game in a row that puts the 5600 XT boosted-clocked version ahead of the RTX 2060. The stock-clocked version slides in just behind it. What a difference a BIOS can make!
On the previous page, we saw the 5600 XT give the RTX 2060 a great fight in every test, but NVIDIA manages to strike back in Monster Hunter World, with it, and the GTX 1660 SUPER soaring to the top. Seeing the 1660 SUPER outpace a card like the RX 5700 is really intriguing, considering the SRP on the 5700 is around $349, and the 1660 SUPER, $229.
Clearly, optimizations are going a long way for NVIDIA with this title, so we’d lean towards that being the better choice if you take this game really seriously (and it seems like many do).
Since its launch, MHW has dramatically upgraded itself to support newer technologies. The game didn’t even launch with ultrawide support, but that was taken care of fairly quickly. Since then, a high-res texture pack has also come out, and support for both NVIDIA’s DLSS and AMD’s FidelityFX have been added – so regardless of whether you go AMD or NVIDIA, you have some extra toys to play with.
With SotTR, AMD’s boosted-clock RX 5600 XT matches NVIDIA’s RTX 2060, while the 5600 XT reference clock of course falls a bit behind. Well, “a bit” being 9 full frames on average. It should be clear by this point that if you want the best-performing 5600 XT, you’ll definitely want to make sure you get one with the increased clocks, which includes the 14Gbps memory.
This is another title where the RX 5600 Pulse concerns the RX 5700 just a wee bit. The difference between RX 5600 XT reference and Pulse is 9 FPS, and that happens to be the same difference between the Pulse and RX 5700.
Three Kingdoms has been promoted as working on Intel’s integrated graphics – it even has an Intel logo while the game starts up. That’s all fine and good, but the reality is, Three Kingdoms, like most Total Wars, has an obscene number of dials that can be turned, so no matter the hardware you’ve got, you should be able to reach playable framerates easily, or at least eke out a ton of graphical IQ.
A result that stands out here is between the two RX 5600 XT configurations. The reference card hit 72 FPS at peak, which happened to be the Sapphire Pulse’s minimum, with the overall advantage being 8 FPS on average. Another interesting result is that the RX 590 performs the same as RX 5500 XT, despite the latter being the much newer replacement.
High-FPS gaming generally loves faster memory, which is one of the reasons the 1660 SUPER can sometimes outperform the technically superior 1660 Ti. In the 5600 XT’s case, the 14Gbps memory put the Pulse card at the top of both the 1080p and 1440p results. At 1080p, the scaling is rather boring, and overall, it appears all of those GPUs could handle 144Hz (we’d still not go with a low-end card if competitive gaming is important).
It’s not the focus of this review, but since we’ve brought up the 5500 XT a few times, it definitely should be here. Whereas that card doesn’t always manage to beat out the RX 590, it easily managed it using both resolutions here, and by a rather significant margin. Despite having a lower bandwidth rating, the performance goes from 8Gbps on the RX 590 to 14Gbps on the RX 5500, so we’re seeing clear advantages from faster memory here.
At both tested resolutions in Siege, we’re seeing a massive difference between the reference-clocked RX 5600 XT and the Sapphire Pulse boosted one. Again, we’re dealing with high-FPS here, but if your goal is to maximize your 144Hz monitor, you’ll obviously want to make sure you have as much extra breathing room as possible. A 148 FPS average at 1440p doesn’t mean it’ll never dip below 144Hz, so you have to keep that in mind (we haven’t recorded minimums here because they haven’t been consistent enough).
In CS: GO, the RX 5500 XT outpaced the RX 590 by quite a bit, but the situation changes in Siege, where the older RX 590 actually manages to be the better-performing card. It could be that this game just happens to be really optimized for the Polaris architecture, so hopefully the newer RX 5500 XT can take the lead here at some point in the future.
While many compare against just one 3DMark score, it’s important to look at the bigger picture. In this case, that means running both the DX11 and DX12 tests to see how things change up. On the RX 5600 XT front, the Sapphire Pulse retail card boosts performance by about 13% in the DirectX 11 test, and 10% in DirectX 12. If you needed any more proof that you want the boosted-clock RX 5600 XT, this is it.
In Fire Strike, the RX 5600 XT Pulse outperforms even the RTX 2060 SUPER – not just the regular RTX 2060. Things change with the DX12 Time Spy, with even the boosted Pulse card falling behind the original RTX 2060. See? One single set of results would hide these scaling differences.
How does all that compare to real-world gaming? Well, we’ve had a bunch of that on the previous pages. 3DMark puts DX12 performance in NVIDIA’s favor, but we have not consistently seen that behavior in the real-world. In Borderlands 3, for example, we’ve found AMD’s DX12 to perform better than NVIDIA’s – but that’s just a single game.
VRMark is similar to 3DMark in that multiple tests could deliver different scaling. In the Cyan room test, the 5600 XT Pulse proves to be much better than the reference clock, beating out the RTX 2060 by a few frames. In the Blue room, the roles again change a bit, with NVIDIA having an overall advantage. Obviously, as you can tell by the FPS results here, the absolutely only GPU that passes the test is the RTX 2080 Ti. Thankfully, this Blue room test represents future workloads. The more modest Cyan room gives far better results if translated to real VR use.
To wrap up our testing, we have Superposition. Mimicking what we’ve seen a bunch of times before, the Sapphire Pulse clocks outperform the reference clocks by a fair margin. Again, in one test, 1080p, the RTX 2060 outperforms the Pulse, but a trade is made at 4K, which uses less extreme settings, but at a higher resolution.
It’s a little difficult to draw conclusions for AMD’s Radeon RX 5600 XT. The BIOS confusion hasn’t helped matters at all, and we feel that there is still a lot of confusion lingering. We were provided the clock-boosting BIOS ahead of embargo, which we assumed meant the old data could be trashed. Instead, that old data is more relevant than that from the updated BIOS.
Despite the fact that the Sapphire Pulse we used for our testing supports the boosted clocks, including 14Gbps memory, the card is still listed as 12Gbps at sites like Amazon. We assume that this is because the specs can’t be claimed to be higher when they’re not, out-of-the-box. We foresee many buying this card, not even realizing that a simple BIOS flash would increase their performance by ~10%.
The obvious match-up throughout this article has been between the RX 5600 XT and NVIDIA’s GeForce RTX 2060. Depending on whether you’re using the reference or updated BIOS, the 5600 XT will slot into performance charts differently. With reference clocks, the RX 5600 XT only surpassed RTX 2060 performance in one of our eight games. With the upgraded BIOS, that becomes six of eight. It could be said that the BIOS update was a counter to NVIDIA dropping the price of the RTX 2060, but leaving whether the update should be rolled out or not, up to the AIBs.
Over at Amazon once again, the RTX 2060 can be had for as low as $320, which means a $30 premium over this particular Sapphire card. Considering the BIOS upgrade puts it ahead in gaming more often than not, that gives the RX 5600 XT a good impression of value. That said, we cannot even guarantee that your RX 5600 XT will have those updated clocks, so we don’t entirely feel safe summing up a single card that has two completely different performance profiles.
If you can find a 14Gbps-capable RX 5600 XT for less than the RTX 2060, then it’s a great value. NVIDIA-exclusive perks you’d be missing out on include anything that uses the Tensor and RT cores, either in gaming or creative use. We didn’t test power or temperatures this go-around, but the RX 5600 XT didn’t strike us as particularly loud. We can’t really say the same thing about the RX 5700/XT.
Overall, the RX 5600 XT is a solid release, and priced right. That “priced right” turns into a great value if a faster BIOS can be upgraded to. Buying a specific model GPU shouldn’t be this challenging, but with the RX 5600 XT, it is. If you want one, keep your eyes peeled for the faster models.
Copyright © 2005-2020 Techgage Networks Inc. - All Rights Reserved.