Date: November 15, 2018
Author(s): Rob Williams
AMD’s latest Polaris-based gaming graphics card has arrived. It hits us in the form of the Radeon RX 590, a die-shrunk version of the RX 580 that enjoys a monstrous clock boost. While there isn’t a lot to be surprised about with this release, AMD proves that it hasn’t eased its aggressiveness, having augmented the launch with a killer game bundle.
Here’s a product I never thought would exist: AMD’s Radeon RX 590. Sure, the card has been floating around the rumor mill for a while, but remember when an NVIDIA GeForce card based on Volta also was? Sometimes, rumors just fall flat. And given that the Polaris architecture has already been stretched so much with the RX 500 series, I personally never expected to see a 590.
As much as I never expected this product to exist, though, it doesn’t mean that it shouldn’t. This is in effect a clock-boosted product (RX 570~590 comparison here), but it’s honestly a great clock boost, and does manage to separate the 590 from the 580 a good amount overall. AMD is suggesting pricing of $279 USD for its RX 590, which competes closest with NVIDIA’s GeForce GTX 1060, which has models for $250~$290 on Amazon. The RX 580 goes for about $250 on average, but can currently be found on sale for a mere $199 on Amazon (more on this in the final thoughts).
So how does AMD manage to crank an already nicely clocked RX 580 to 11 with another 200 or so MHz? Well, the transition to a 12nm process sure can’t hurt, and while that isn’t likely the sole reason for the boost, it makes me really want to see RX Vega built on a smaller process. That tasty new 7nm from TSMC that AMD has been trialing can’t get here quick enough.
AMD handled full sampling for the RX 590, so I can only imagine that I was chosen to receive a “Fatboy” model for an obvious reason – one that has nothing to do with Harley-Davidson. The reason for the card to be called Fatboy is because it takes up more than 2 slots, but only slightly. I personally do not think this is justified for a mid-range product, but I’m not going to judge when the GPU can hit 78°C with ease as-is.
Aside from the reminder that I need to purge some pounds, XFX’s Fatboy can take advantage of Radeon Chill, to run super-quiet when the load is not that great, includes a backplate for improved heat dissipation, and has a dual-BIOS for those who like to get crafty with their gear. The card is also said to be overclockable (by XFX), but that’s testing I skipped since I had to catch up on 1080p benchmarks for this article (since I had none from the last suite update).
|AMD’s Radeon Gaming GPU Lineup|
|Cores||Base MHz||Peak FP32||Memory||Bandwidth||TDP||Price|
|Vega 64||4096||1546||12.6 TFLOPS||8 GB 4||484 GB/s||295W||$499|
|Vega 56||3584||1471||10.5 TFLOPS||8 GB 4||410 GB/s||210W||$449|
|RX 590||2304||1576||7.1 TFLOPS||8 GB 3||256 GB/s||225W||$279|
|RX 580||2304||1340||6.1 TFLOPS||8 GB 3||256 GB/s||185W||$229|
|RX 570||2048||1244||5.1 TFLOPS||8 GB 3||224 GB/s||150W||$179|
|RX 560||896||1175||2.6 TFLOPS||4 GB 3||112 GB/s||80W||$119|
|RX 550||640||1183||1.2 TFLOPS||2 GB 3||112 GB/s||50W||$99|
|Notes||1 GDDR6; 2 GDDR5X; 3 GDDR5; 4 HBM2|
Architecture: RX 550~590 = Polaris; RX Vega 56 & 64 = Vega
When PowerColor decided to leak the RX 590 earlier this week, its website reported on a 17% performance boost. The RX 590 is rated at 7.1 TFLOPS by AMD, which gives us about a 16% boost over the RX 580. I’d tell you if that proves true right here if I didn’t have a couple of pages ahead I wanted you to actually look at. But, you can probably take an accurate guess if you don’t feel like clicking ahead.
There is really not too much else to say, so let’s get going:
|Techgage Gaming GPU Test PC|
|Processor||Intel Core i7-8700K (3.7GHz, 6C/12T)|
|Motherboard||EVGA Z370 FTW|
CPU tested with BIOS 1.09 (August 20, 2018)
|Memory||G.SKILL TridentZ (F4-3400C16-8GSXW) 8GB x 2|
Operates at DDR4-3400 16-16-16 (1.35V)
|AMD Graphics||AMD Radeon RX Vega 64 (8GB; Radeon 18.8.2)|
AMD Radeon RX Vega 56 (8GB; Radeon 18.8.2 + Nov 14)
AMD Radeon RX 590 (8GB; Radeon Nov 14)
AMD Radeon RX 580 (8GB; Radeon 18.8.2 + Nov 14)
AMD Radeon RX 570 (4GB; Radeon 18.8.2 + Nov 14)
AMD Radeon RX 550 (2GB; Radeon 18.8.2)
|NVIDIA Graphics||NVIDIA GeForce RTX 2080 Ti (11GB; GeForce 411.51)|
NVIDIA GeForce RTX 2080 (8GB; GeForce 411.51)
NVIDIA GeForce RTX 2070 (8GB; GeForce 416.34)
NVIDIA TITAN Xp (12GB; GeForce 399.07)
NVIDIA GeForce GTX 1080 Ti (11GB; GeForce 399.07)
NVIDIA GeForce GTX 1080 (8GB; GeForce 399.07)
NVIDIA GeForce GTX 1070 Ti (8GB; GeForce 399.07)
NVIDIA GeForce GTX 1070 (8GB; GeForce 399.07 + 416.34)
NVIDIA GeForce GTX 1060 (6GB; GeForce 399.07 + 416.34)
NVIDIA GeForce GTX 1050 Ti (4GB; GeForce 399.07)
NVIDIA GeForce GTX 1050 (2GB; GeForce 399.07)
|Storage||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; build 17134)|
Since 1080p testing hadn’t been conducted at all since the last test machine overhaul a few months ago (which will soon be overhauled again with the i9-9900K), the GTX 1060, GTX 1070, RX 570, RX 580, RX 590, and Vega 56 were tested with the latest driver only at that resolution. While 399.07 to 416.34 seems like a mammoth jump, it actually represents only two months of driver releases.
A total of eight games are included in our current test suite. Some have appeared here before, while others make their first appearance: Monster World Hunter, Fortnite, and F1 2018. I had planned to include PlayerUnknown’s Battlegrounds as a ninth title, but the results were too sporadic to inspire any sort of confidence (an issue not seen in Fortnite, by comparison).
Here’s the full list of tested games and developer allegiances, as well as synthetic tests also used:
This article is going to be the last for this particular suite, as some games are now on the market that should be slipped in here. Battlefield 1 will be replaced with Battlefield V, and Rise of the Tomb Raider will be replaced with Shadow of the Tomb Raider. I might also drop Fortnite since there hasn’t been much demand, and it’s a time-consuming chore to test (a built-in benchmark for online games would be great).
For our apples-to-apples testing, the graphics settings seen above apply to every one of our tested resolutions so as to deliver standard apples-to-apples results. In most cases, each configuration is tested twice, with more runs added if the initial results make the extra testing necessary (which is actually rare, even with manual runs). Fortnite is the only game tested three times by default due to its variable nature.
While it is not at all targeted at the resolution, I ran the RX 590 through 3440×1440 tests since it was easy to do, and I wanted to see if it was possible to get any reasonable framerates out of the wider 1440p resolution. 4K has about 70% more pixels than 3440×1440, so… don’t count on RX 590 being too useful there. Ideally, the more modest ultrawide resolution of 2560×1080 would have been tested, but time unfortunately didn’t allow it.
Similarly, I also tested the Vega 56 at 1080p only because I saw some comments around the web requesting the comparison. Apparently, Vega 56 can sometimes drop to nearly $300 (though I have not yet seen it), so if you ever find a blockbuster deal like that, you don’t even think about it – you just splurge the extra, and forget about the RX 590.
The higher the resolution, the smaller the gains the RX 590 will exhibit over the RX 580. At 1080p, AMD’s latest gains 7 FPS, which drops to a gain of 5 FPS at 1440p. As I analyze these results, I now realize that I didn’t test the GTX 1060 in the higher resolutions, so I’m glad it’s at least covered in the 1080p set. AMD really cleans house at 1080p, though, slotting the three RX cards right in between the 1060 and 1070.
The RX 590 continues to leap ahead of the RX 580 by a fairly expected amount, again sitting in front of NVIDIA’s GTX 1060 – and not at all far behind the GTX 1070.
At 1440p, the RX 590 with our settings can handle about 50 FPS, which isn’t bad at all given the level of detail. However, that consists of a 40 FPS minimum, which might be a bit low for your tastes. Fortunately, Mankind Divided offers a ton of tweaking ability, so you won’t have a hard time getting a 10 FPS boost.
In F1 2018, the GTX 1060 manages to overtake the RX 570 at 1080p, and once again, the RX 590 slots in ahead of the RX 580 with a 7 FPS increase. While Mankind Divided had tight performance between the RX 590 and GTX 1070, the latter card clearly spreads its wings better here (and with Battlefield 1, but we have many more benchmarks to go).
These 7 FPS increases for the RX 590 at 1080p over the RX 580 seem to be rather consistent. Yet again, the GTX 1060 displaces the positioning of the RX 570, after which point the cards scale as expected.
To the surprise of no one, I’m sure, the RX 590 can handle Fortnite just fine at 1080p. And 1440p, for that matter. Though if you are insistent on higher than 60 FPS in your online shooters, you can easily drop the detail levels down and gain an instant boost. Interestingly, this is the first test where the delta between the RX 580 and 590 is quite small – a mere 3 FPS.
Due to the nature of an online game, it’s difficult to get tight performance deltas from run to run. With mega-popular games like these, it’d be nice if the developer could cough up a worthwhile in-game benchmark that actually allows true apples-to-apples testing. That said, Fortnite is actually pretty consistent, all things considered, whereas PUBG in my experience has been anything but. 20 FPS deltas from run to run in that game were common, which is why it isn’t used in this lineup.
That tangent aside, this is the first example of a title where NVIDIA has some clear optimizations, as its 4.3 TFLOPS GTX 1060 outperformed the 7.1 TFLOPS RX 590. Likewise, the GTX 1070 outperformed the technically faster RX Vega 56.
Monster Hunter World doesn’t strike me as an optimized game, but as mentioned in previous articles, it scales as expected, and since an unoptimized game still represents current performance, it remains relevant.
Here, the RX 590 jumps 5 FPS ahead of the RX 580 at 1080p, with the GTX 1060 once again displacing the RX 570 for the second-from-bottom positioning.
The beefier cards exhibit some huge strengths with Rise of the Tomb Raider, but at 1080p, all of the cards here deliver better than acceptable performance. At 1440p, no real compromises have to be made. At max detail, the RX 590 delivers a clean 60 FPS on average.
Wildlands came out before Far Cry 5, but in many ways, I consider Wildlands to be the much more beautiful game. It’s also one I suggest anyone who enjoys open-world shooters dive into. There’s a lot of content, and the game only gets better with friends in co-op. But I digress.
At 1080p, all of the cards can handle this game without issue at max detail, with the RX 590 keeping comfortably ahead of 60 FPS, and coming close to that for the minimum. Even 1440p performance is good, though I’d personally tweak something to inch a bit closer to 60 FPS.
While there were some shakeups in some of the game tests, the DirectX 11 Fire Strike has shown the exact same kind of scaling between the GTX 1060, 1070, and RX 570~590 as we’d expect. According to 3DMark, the RX 590 is about 10% faster than the RX 580. The same fact remains for the DirectX 12 Time Spy:
In the DirectX 11 test, the GTX 1070 jumped 14% ahead of the RX 590, but in this DirectX 12 test, that increase jumps to 25%. Likewise, the GTX 1060 which fell behind the RX 570 in the DX11 test has reversed roles in the DX12 one. NVIDIA’s Turing architecture extends that dominance to the top of the chart.
In the Blue Room test, which represents future VR workloads, the RX 590 struggled a bit. This is the first test outside of Fortnite where the GTX 1060 competes head-to-head with the RX 580. Clearly, this is an extremely tough workload, and the performance deltas are going to be small because of that, but clearly, NVIDIA has an advantage with this test. And look at that RX 550 result – ouch.
With Superposition, it seems like the only advantage a GPU will have is its own strength. The cards scale pretty much as I’d expect, outside of the RTX 2080 Ti keeping far ahead of the rest. Whereas 3DMark suggested a 10% performance gain in the DX11 Fire Strike test, Unigine settles on that same value for the 4K test, but sees a slight increase to 12% in the 1080p test.
For its gain in performance, the RX 590 uses about 30W more than the RX 580. That’s a bit of a jump, but it’s not an unexpected one given the huge clock boost. That 30W increase is still less than the 40W jump AMD itself reports (185W vs. 225W), so in that regard, we seem to be coming ahead. Not that power consumption is a major issue, but it is almost painful to see the faster GTX 1070 use 96W less.
There isn’t too much I can say here that isn’t obvious. The RX 590 is faster than an RX 580 as its huge ~200MHz clock boost would suggest. What your purchase will likely ultimately boil down to is how much you want to spend, and what kind of performance you’re after. With a price tag of $279, I’d say the RX 590 is priced right, but there are some things to bear in mind.
While the RX 590 currently retails for $279, the RX 580 is currently on sale for $199 at both Newegg and Amazon. The timing of this sale is actually crazy to me. How would I recommend a $279 GPU that gives a 3~7 FPS boost over an RX 580 that currently costs $199? A $20 premium is worth it, but not an $80 one.
It gets better, because the RX 570 and RX 580 qualify for the game bundle that launches alongside the RX 590. That means that you can pick up an RX 580 right now for $200 and enjoy two of the three games in the bundle. It’s a weird way to look at free games, but if you were to buy those games anyway, that’d make the card cost $80 (based on $60 per title).
Two games and RX 580 for $199 sounds more tempting to me right now than three games and RX 590 for $279. But that’s just me. This $199 pricing is dubbed a sale over at Newegg, so it’s not going to last. Maybe there was no good time to put that card on sale, but really – for $199, the RX 580 is a steal. Especially with two free games.
That said, you’ll want to of course verify that your card purchase will qualify for this bundle. You can read full details of it right here. RX 590 owners will get Resident Evil 2, Devil May Cry 5, and Tom Clancy’s The Division 2 – while RX 570 and RX 580 purchasers can choose two. Honestly, this is one of the best game bundles I’ve seen, so I give AMD props for putting it together. If there’s a downside, it’s that all three games won’t launch until early 2019.
tl;dr: Pricing is everything. Scout out what gives you the best ROI, or simply jump for the big gun of the Polaris world and get on with life.
When all is said and done, I rate the RX 590 pretty much the same as the RX 580 in our expectations chart. It does offer a noticeable boost, but not enough to justify any changes here. I debated on adding a star for 1440p, but at that point, it’d match the GTX 1070, when the reality is different.
Game Performance Expectations
|RTX 2080 Ti||★★★★★||★★★★★||★★★★||★★★★||★★★★|
|GTX 1080 Ti||★★★★★||★★★★★||★★★★||★★★||★★★★|
|RX Vega 64||★★★★★||★★★★||★★★||★★||★★★|
|GTX 1070 Ti||★★★★||★★★||★★||★||★★|
|RX Vega 56||★★★★||★★★||★★||★||★★|
|GTX 1050 Ti||★★||★||★||★||★|
|144Hz values based on 1080p with high detail (independent of the logic below:)|
★★★★★ 60 FPS? More like 100 FPS. As future-proofed as it can get.
★★★★ Surpass 60 FPS at high quality settings with ease.
★★★ Hit 60 FPS with high quality settings.
★★ Nothing too impressive; it gets the job done (60 FPS will require tweaking).
★ Not recommended.
If you have questions not covered here, please shoot them at me. Overall, this is a good release from AMD, as much as it is an odd one in a way. Still, more choice in the market is a great thing, and when AMD is offering a killer game bundle as it is… the launch becomes quite alluring.
Copyright © 2005-2018 Techgage Networks Inc. - All Rights Reserved.