Date: February 22, 2019
Author(s): Rob Williams
NVIDIA’s GeForce GTX 1660 Ti becomes the fifth card based on Turing to be released, but unlike the RTX cards we’ve seen up to this point, the 1660 Ti foregoes Tensors and RT cores in favor of delivering a more competitively-priced product, and an all-around enticing competitor. Let’s see how it stacks up against NVIDIA’s own lineup, and AMD’s competition.
The second GeForce release to come out in 2019 becomes the GTX 1660 Ti, a Turing-based card that’s just a bit different from the others (as its GTX moniker implies). Gone are the Tensor and RT cores, but retained are the other performance perks that came along with Turing, leading to some surprising performance in select games and benchmarks.
NVIDIA dropped RTX features off of the GTX 1660 Ti for the simple fact that the card is not powerful enough to take proper advantage of those features. We were a bit surprised when the 2060 came out with RTX features, but it ultimately delivered enough performance to justify them.
Key features Turing brings to the table, even to the GTX 1660 Ti, includes concurrent FP/INT operations, benefiting games that heavily use both floating-point and integer operations. Far Cry 5 and Shadow of the Tomb Raider can use a lot of both in certain areas of the games. NVIDIA notes that in SotTR, 38% of the game’s instructions are integer on average, and lo and behold, the 1660 Ti performs great there.
Turing also supports variable rate shading, allowing games to take advantage of adaptive shading to reduce load in unimportant parts of a scene, putting most of the GPU horsepower towards what will actually prove beneficial to the scene. So far, Wolfenstein II: The New Colossus takes advantage, and we have no choice but to wait to learn about more.
The GTX 1660 Ti includes GDDR6 just like the RTX cards, and it’s spec’d at an impressive 12Gbps. With this TU116 design, NVIDIA delivers 288GB/s of memory bandwidth, which is 50% higher than the GTX 1060’s 192GB/s.
|NVIDIA’s GeForce Gaming GPU Lineup|
|Cores||Base MHz||Peak FP32||Memory||Bandwidth||TDP||SRP|
|TITAN RTX||4608||1770||16.3 TFLOPS||24GB 1||672 GB/s||280W||$1,199|
|RTX 2080 Ti||4352||1350||13.4 TFLOPS||11GB 1||616 GB/s||250W||$999|
|RTX 2080||2944||1515||10.0 TFLOPS||8GB 1||448 GB/s||215W||$699|
|RTX 2070||2304||1410||7.4 TFLOPS||8GB 1||448 GB/s||175W||$499|
|RTX 2060||1920||1680||6.4 TFLOPS||6GB 1||336 GB/s||160W||$349|
|GTX 1660 Ti||1536||1500||5.5 TFLOPS||6GB 1||288 GB/s||120W||$279|
|TITAN Xp||3840||1480||12.1 TFLOPS||12GB 2||548 GB/s||250W||$1,199|
|GTX 1080 Ti||3584||1480||11.3 TFLOPS||11GB 2||484 GB/s||250W||$699|
|GTX 1080||2560||1607||8.8 TFLOPS||8GB 2||320 GB/s||180W||$499|
|GTX 1070 Ti||2432||1607||8.1 TFLOPS||8GB 3||256 GB/s||180W||$449|
|GTX 1070||1920||1506||6.4 TFLOPS||8GB 3||256 GB/s||150W||$379|
|GTX 1060||1280||1700||4.3 TFLOPS||6GB 3||192 GB/s||120W||$299|
|GTX 1050 Ti||768||1392||2.1 TFLOPS||4GB 3||112 GB/s||75W||$139|
|GTX 1050||640||1455||1.8 TFLOPS||2GB 3||112 GB/s||75W||$109|
|Notes||1 GDDR6; 2 GDDR5X; 3 GDDR5; 4 HBM2|
Architecture: GTX & TITAN = Pascal; RTX = Turing
NVIDIA says that against the GTX 1060, the 1660 Ti can give us significant performance boosts, upwards of 50%. That alone means that it’s worthy of consideration by those who own a 1060, because a 50% gain is no joke. But compare it to the GTX 960, which came out in 2015. In that match-up, the 1660 Ti is about 2.2x faster, so if you are still rocking that one, this 1660 Ti would be a tempting upgrade path.
Admittedly, it might be safe to remove some 10-series cards from this table, but for the sake of showing what both generations offer, they’re useful to keep for now, even if it’d be silly to purchase a 10-series card when an equivalent 20-series card is available (and now 16-series…)
The 1660 Ti slots in rather close to the RTX 2060, which justifies its $70 premium with some extra performance, and the inclusion of Tensor and RT cores. Whether or not that premium is worth it is really up to you, but it’s nice to see the price gap isn’t that extreme between them.
And with that, we’ll get into a quick look at our test methodology, and then jump on the test results.
|Techgage Gaming GPU Test PC|
|Processor||Intel Core i9-9900K (3.6GHz Base, 5.0GHz Turbo, 8C/16T)|
|Motherboard||ASUS ROG STRIX Z390-E GAMING|
CPU tested with BIOS 0602 (October 19, 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 18.12.3)|
AMD Radeon RX Vega 56 (8GB; Radeon 18.12.3)
AMD Radeon RX 590 (8GB; Radeon 18.12.3)
AMD Radeon RX 580 (8GB; Radeon 18.12.3) *
AMD Radeon RX 570 (4GB; Radeon 18.12.3) *
AMD Radeon RX 550 (2GB; Radeon 18.12.3) *
|NVIDIA Graphics||NVIDIA GeForce RTX 2080 Ti (11GB; GeForce 417.35)|
NVIDIA GeForce RTX 2080 (8GB; GeForce 417.35)
NVIDIA GeForce RTX 2070 (8GB; GeForce 417.35)
NVIDIA GeForce RTX 2060 (6GB; GeForce 417.35)
NVIDIA GeForce GTX 1660 Ti (6GB; GeForce 418.91)
NVIDIA TITAN Xp (12GB; GeForce 417.35) *
NVIDIA GeForce GTX 1080 Ti (11GB; GeForce 417.35)
NVIDIA GeForce GTX 1080 (8GB; GeForce 417.35) *
NVIDIA GeForce GTX 1070 Ti (8GB; GeForce 417.35) *
NVIDIA GeForce GTX 1070 (8GB; GeForce 417.35)
NVIDIA GeForce GTX 1060 (6GB; GeForce 417.35)
NVIDIA GeForce GTX 1050 Ti (4GB; GeForce 417.35) *
NVIDIA GeForce GTX 1050 (2GB; GeForce 417.35) *
|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 17763)|
|Notes||* Synthetic tests only.|
Since the time AMD’s GPUs were originally tested, newer drivers have released, such as the latest: 19.2.2. We retested the RX 590 at 1080p using that 19.2.2 driver, and found no difference in performance. We also tried to run the same tests again on Vega 56, but couldn’t get past the driver install in time.
A total of eight games are included in our current test suite. Recent additions include Battlefield V, Forza Horizon 4, and Shadow of the Tomb Raider. Beyond these eight titles, UL’s 3DMark and VRMark, as well as Unigine’s Superposition, are used for some quick and dirty tests that you may be able to run at home.
Here’s the full list of tested synthetic benchmarks, games, and developer allegiances:
For our apples-to-apples testing, the graphics settings seen above apply to every one of our tested resolutions so as to deliver easily comparable results. In most cases, each configuration is tested twice, with more runs added if the initial results make the extra testing necessary (which isn’t required too often). Note that VSync is disabled at the driver level to prevent games from enabling it without us noticing.
Battlefield V starts us off with a 1660 Ti that’s very competitive with AMD’s Radeon RX Vega 56. Given the $279 pricing of the 1660 Ti, it’s easy to assume it’d perform similarly to the $279 RX 590, but there’s a decent 11 FPS average gap at 1440p, a lead that grows to 14 FPS at 1080p.
It seems strange that this $279 card keeps up to the Vega 56, but it’s not something that’s a surprise to AMD. Last night, we were shot over an email from the company talking of a $279 Vega 56 over at Newegg, which of course was sold out by the time anyone clicked through. There’s no telling if that kind of pricing is going to become static, but if it does, the RX 590 would then need a drop to reflect the gap in performance at the same price point. We’ll see what happens.
In case anyone asks, there are in fact a few models missing from the 1440p chart that could be included. That includes the GTX 1080, and possibly the 1070 Ti. The results seen above were merged from both our RTX 2060 and Radeon VII reviews, and certain cards, like the 1080, were not tested for either of those. We received our 1660 Ti less than 24 hours before embargo, and didn’t realize when the embargo lift even was, so there was no time to get other cards tested in time.
In Battlefield V, the 1660 Ti pulled a fair bit ahead of the RX 590, while almost matching the RX Vega 56. In Deus Ex, the battle of the $279 cards becomes closer, but NVIDIA still ekes about 5 FPS more on average.
Meanwhile, the Vega 56 outperforms the 1660 Ti by a fair margin here, which makes you realize if that model did become available regularly for $279, it’d offer a ton of value. That’s not something we can count on, though, especially since the least-expensive Vega 56 we can find as of the time of writing on Newegg is $400 ($350 if you want to go refurbished).
AMD might be the company with its logo on Formula 1 cars, but it’s NVIDIA that dominates in F1 2018. Here, the 1660 Ti outperformed the Vega 56 at 1440p, and matched it at 1080p. Thanks to this, the gain on 1660 Ti over RX 590 is significant – about 15 FPS at 1440p, and 22 FPS at 1080p.
It feels like AMD is a big part of this review even though its focus is on an NVIDIA card, because we continue to see the 1660 Ti give the RX 590 a hard time, and remember, that card only released in November. NVIDIA really didn’t give AMD much time to enjoy the $279 price point before charging in with this oddly-named GTX. The RX 590 already felt like a good value, so by default, the 1660 Ti feels like a good value by this point – but, we still have more tests to conquer.
FC5 is an example of a game that happens to work better on hardware opposite of the game sponsor. Far Cry 5 was heavily promoted as being best played on Radeon, yet here we are, with NVIDIA exhibiting some notable leads.
AMD typically performs very well in Forza games, having promoted both Forza 7 and Horizon 4 for benchmarking to some extent, but sadly, the #BetterRed team falls far behind NVIDIA here. The RX Vega 56 is somehow 4 FPS slower than the 1660 Ti at 1440p, a gap that widens to 9 FPS at 1080p. Naturally, edging out Vega 56 means that the 1660 Ti really outperforms the RX 590 here.
We’ve come to yet another title that shows very strong performance on the 1660 Ti against Vega 56, with the lowbie NVIDIA card once again beating that competition out at 1080p. At 1440p, the going gets a little more tough, and the RX Vega 56 begins to strut its stuff, placing 2 FPS ahead. Still, that leaves the 1660 Ti well ahead of the RX 590, which is its actual competition at the $279 price point.
Yet again, we’re seeing super strong performance from the 1660 Ti in its battle with the Vega 56 – a battle we were not even considering when testing. To us, the competition has been the RX 590, yet that card is falling behind a fairly significant margin often. Granted, we’re only testing with eight game titles here, but there’s an even spread between vendor-favoring titles between them, and yet NVIDIA is still coming ahead of the RX 590 each and every time.
We wrap up our testing with another example of the 1660 Ti outperforming the RX 590 competition which only released in November. NVIDIA clearly has an answer for anything AMD wants to put out, which puts the red team in an awkward position. Given this performance, it’s going to be hard to sell an RX 590 for $279 when NVIDIA’s $279 option regularly goes toe-to-toe with the even higher-end RX Vega 56.
In our real gaming tests, we saw the 1660 Ti outperform the RX 590 in many cases, and keep up to the Vega 56 just as often. In multiple cases, the 1660 Ti even managed to outperform the Vega 56, which at SRP costs much more. But now we come to 3DMark, which paints a different picture. Here, the V56 still pulls far ahead of the 1660 Ti, especially at 4K. With Time Spy‘s 4K test, optimizations help push the 1660 Ti closer towards the Vega 56.
Time Spy, being DX12 based, seems to run extremely on Turing, which now includes the GTX 1660 Ti. Gains there could be partly fueled by NVIDIA’s ability to run concurrent FP and INT operations. Shadow of the Tomb Raider makes use of that, and lo and behold, the 1660 Ti manages to keep up to the Vega 56 there, as well.
We saw the 1660 Ti come close to RX Vega 56 performance in 3DMark’s Time Spy test, and in VRMark, we see similar things. All of the Turing-based GeForces sit near the top of the charts, with the newer RTX 2060 matching the last-gen GTX 1080 in the Cyan room, and matching it in Blue.
Wrapping up our performance testing, Superposition shows the 1660 Ti falling behind the RX Vega 56 a little bit at 1080p, but the 4K test makes the cards almost equal. It’s clear that Turing has some very useful optimizations, easily seen in some benchmarks.
To test for power consumption, a Kill-A-Watt that the PC itself is plugged into is used for monitoring a Far Cry 5 4K benchmark run. Admittedly, we don’t have the best methods for power testing, and would love to improve them in the future, but for now, we use what we have, and that’s our eyeballs. Over the course of the one-minute benchmark, a rough average is pulled.
With its performance, the 1660 Ti still manages to sip a modest amount of power compared to some of the higher-end cards. Despite not being far off from Vega 56 performance in many cases, that card draws 60W more at load, which is actually still less than what the RX 590 draws (despite that card being slower).
On the noise front, it’s hard to imagine one 1660 Ti is going to be much worse than another, but in the case of this EVGA Xc card, it runs fairly quiet even when the going gets tough. We don’t measure sound, but it’s not hard to “hear” a card even if you don’t want to, so the loud cards always stand out, and this one certainly didn’t.
NVIDIA’s GeForce GTX 1660 Ti didn’t quite come out of nowhere, as it’s been a major subject of interest in the rumor mill for a while. However, the name itself hasn’t been really solidified until recently. There were previous rumors about this card being called 2050 Ti. It seems likely that NVIDIA chose the 10-series esque name on purpose, simply to avoid confusion with regards to RTX features. Not that 1660 isn’t a strange name in itself, but so was GTX 1080 when it came out, and we all seemed to get used to it.
Against AMD’s Radeon RX 590, the GTX 1660 Ti cleaned house in every single one of our real game tests, and that alone is enough to tell you which card you should be seeking out. This is unfortunate for AMD, since the 590 only released in November, and the 1660 Ti isn’t exactly kind with its leads sometimes.
As mentioned in the intro, AMD reached out to us last night to tout a $279 Vega 56 that could be found at Newegg. These kinds of emails ahead of a competitor launch are common, and perhaps more common is the fact that the deals are gone before the vast majority of fans have a chance to take advantage. AMD wanted to throw a wrench in NVIDIA’s gears a bit ahead of this launch, and it seems unlikely that we’ll see many more $279 Vega 56s hit etail in the near-future.
That leads us to the RX 590. It could still be competitive with a price drop, but the sad fact is, this was a card that was competitive right on up until this 1660 Ti launched. AMD will have to be more aggressive with that pricing, especially if it ever wants to drop Vega 56 to $279 again. Doing that wouldn’t benefit AMD, as Vega cards are expensive to make thanks to their inclusion of HBM.
Nonetheless, pricing will dictate everything here, so whatever you’re after, patience will pay off if you want to monitor etailers and see when the best deal drops. As it stands today, though, the 1660 Ti at its $279 price point is impossible to beat. The RX 590 draws 83W more than the 1660 Ti at full load, and even if you say you don’t care about power, 83W is a huge amount for a card of equal or less performance.
At this point, we’ve had our 1660 Ti for under 24 hours, but our first impressions are actually really good, which we admit was kind of unexpected. I was told by an industry friend that I wouldn’t be too impressed, yet I am. It’s hard to not be impressed when NVIDIA is infusing a ton of performance into the $279 price point.
Copyright © 2005-2019 Techgage Networks Inc. - All Rights Reserved.