Nvidia GTX 1660 Ti vs GTX 1070: Which graphics card should you buy?
I know it seems like I’ve been comparing the new Nvidia GeForce GTX 1660 Ti with practically every graphics card under the sun recently, including Nvidia’s RTX 2060 and GTX 1060, but now it’s time to take one final look at how it stacks up against its other main competition, the GTX 1070. Costing just a little bit more than the GTX 1660 Ti at time of writing, is there any hope at all that the GTX 1070 might still be worth buying? Let’s find out with the help of some lovely graphs.
GTX 1660 Ti vs GTX 1070: How we test
Just like my GTX 1660 Ti vs GTX 1060 and GTX 1660 Ti vs RTX 2060 articles, I’ve tested each card with my general games benchmarking suite at 1920×1080 and 2560×1440 resolutions. That includes Assassin’s Creed Odyssey, Shadow of the Tomb Raider, Final Fantasy XV (with Nvidia’s fancy HairWorks, TurfEffects, VXAO and ShadowLibs effects turned off), Monster Hunter: World, The Witcher III, Forza Horizon 4, Total War: Warhammer II and Metro Exodus.
Paired with my Intel Core i5-8600K CPU and 16GB of Corsair Vengeance 2133MHz RAM, I’ve seen what each card can do on max settings at each resolution and taken an average frame rate (either from their own built-in benchmarks or from my own repeated manual gameplay tests) to see how they stack up against each other. In the case of 2560×1440 gameplay, I’ve also looked at what each one is capable on High quality settings, too.
And if you’re wondering why the two graphics cards up the top there look exactly the same, that’s because the cards in question are the Asus’ ROG Strix GeForce GTX 1660 Ti OC and the Asus ROG Strix GeForce GTX 1070 Gaming, showing how little Asus’ ROG Strix design has deviated over the years.
Both are some of the faster varieties of their respective cards out there, so the results below are probably more of a best-case scenario for what each one’s capable of across the different resolutions and graphics settings, but I’m also confident that slower cards will be very much in the same sort of ballpark.
After all, I’ve also been testing Asus’ Phoenix GTX 1660 Ti card recently, and that’s only an average of 2-5fps slower than the ROG Strix (and was more often just 2-3fps slower, too). As a result, I think you’ll be perfectly able to apply most of the results below to the cheaper GTX 1660 Ti and GTX 1070 cards out there as well.
Now we’ve got all that preamble out of the way, let’s dive into those lovely graphs, shall we?
GTX 1660 Ti vs GTX 1070: 1080p performance
On Ultra settings at 1920×1080, both cards are pretty much capable of delivering 60fps right off the bat in a variety of games. The GTX 1070 struggled with Assassin’s Creed Odyssey’s Ultra High settings a bit, and both had a tough time with Metro Exodus’ turned up to eleven (see my Metro Exodus graphics performance guide for a full breakdown of what I consider ‘Ultra’ settings here), but on the whole you’re looking at a smooth minimum of at least 55fps in most cases, making both cards perfectly suited to flawless 1080p play.
However, as you’ll see from the number of taller red bars in the graph above, the GTX 1660 Ti has the upper hand in almost every game going, giving you a boost of around 10fps over the GTX 1070. There are, of course, a couple of games where the GTX 1070 claws back a bit of a lead, such as in Total War: Warhammer II and Metro Exodus, but even here you’re talking about a boost of 1-2fps at best — which, to the naked eye, you’re probably not going to notice.
Indeed, if it’s a graphics card for a high refresh rate monitor you’re after, then the GTX 1660 Ti is definitely the one to go for here as well, as it not only hits 60fps+ more frequently than the GTX 1070 at this resolution, but it can also do it on the best possible graphics setting, too, giving you the best looking games at faster frame rates.
GTX 1660 Ti vs GTX 1070: 1440p performance
Sticking with Ultra settings for a moment, we can see the GTX 1660 Ti is also the superior graphics card at 2560×1440 resolutions, too.
Again, there are a couple of outlying cases where the GTX 1070 either catches up to the GTX 1660 Ti (Assassin’s Creed Odyssey) or just edges it out by another single frame lead again (Metro Exodus), but the rest of the time the GTX 1660 Ti is the clear winner, offering as much as 20fps faster speeds than the GTX 1070.
It is, admittedly, a pretty close race most of the time. In Monster Hunter: World, for example, there’s only an average of 4fps separating each card, and I doubt an average of 54fps in Shadow of the Tomb Raider with its basic SMAA anti-aliasing enabled will feel noticeably different to an average of 62fps. Still, the 8-10fps gap in Total War: Warhammer II and Final Fantasy XV are pretty significant, as far as I’m concerned, as I think almost everyone will agree that a low-40s frame rate is eminently more comfortable to play than something in the low 30s.
The same goes for playing games on High settings at 2560×1440 as well. As you’ll see from the graph below, which mainly focuses on the games that didn’t already result in scores of near-60fps back on Ultra, the GTX 1660 Ti once again takes the lead in practically every game going.
The only exception to that rule is, you guessed it, Metro Exodus, where the GTX 1070 was yet again a single frame in front. I should note that you can still get a smooth 60fps on both cards in this game with High quality enabled, but you’ll need to turn off some of the other special effects in order to make it happen. Monster Hunter: World was also a pretty close call, averaging just 2fps faster with the GTX 1660 Ti, but elsewhere the GTX 1070 was a good 6-9fps behind.
Of course, an average frame rate in the low-to-mid 50s probably isn’t going to be massively different in feel and overall smoothness to one in the high-50s-to-low-60s (at least compared to the low 30s vs low 40s gap on Ultra settings at this resolution), but there’s no denying that the GTX 1660 Ti is a lot more capable of hitting a consistent 60fps here than the GTX 1070. As a result, if you want to be absolutely 100% sure you’re getting 60fps and nothing less, then the GTX 1660 Ti is much more likely to deliver on that promise than its slightly aged relative.
GTX 1660 Ti vs GTX 1070: Conclusion
As you can see, it’s a very closely run race between these two cards, but overall I think you’ll agree that the GTX 1660 Ti is the superior card here. Not only is it more capable of hitting a consistent 60fps in practically every game on 1080p Ultra settings, but it’s also the more reliable card if you’re after 60fps on 1440p High as well.
The GTX 1070 still puts up a pretty good fight here and there, but when so many of its third party cards are more expensive than the £260 / $280 GTX 1660 Ti, it just doesn’t make any sense to go for the GTX 1070 over its newer, cheaper sibling. If prices suddenly take a tumble below the GTX 1660 Ti, then there might be an argument for buying the GTX 1070 in order to save yourself a bit of cash for nearly-similarish sort of speeds, but really, I just don’t think it’s worth it, especially if you’re after a card for playing games on decent quality settings at 2560×1440.
As a result, it’s the GTX 1660 Ti that’s my new best graphics card for 1080p gaming, as well as my ‘What you should actually buy’ recommendation for 1440p gaming if the even faster RTX 2060 is a bit out of your price range (and to see some more graphs on how the GTX 1070 stacks up against the RTX 2060, then check out my RTX 2060 vs GTX 1070 article as well).
GeForce GTX 1660 Ti vs GeForce GTX 1070
- Home
- VGA Benchmarks
- GeForce GTX 1660 Ti vs GeForce GTX 1070
-
GeForce GTX 1660 Ti
101%
-
GeForce GTX 1070
100%
Relative performance
Reasons to consider GeForce GTX 1660 Ti |
| This is a much newer product, it might have better long term support. |
| Supports PhysX |
| Supports G-Sync |
| Supports ShadowPlay (allows game streaming/recording with minimum performance penalty) |
| Supports Direct3D 12 Async Compute |
| Supports DirectX Raytracing (DXR) |
| Supports Deep Learning Super-Sampling (DLSS) |
Reasons to consider GeForce GTX 1070 |
| Supports PhysX |
| Supports G-Sync |
| Supports ShadowPlay (allows game streaming/recording with minimum performance penalty) |
| Supports Direct3D 12 Async Compute |
| Based on an outdated architecture (Nvidia Pascal), there may be no performance optimizations for current games and applications |
No clear winner declared
These graphics cards seems to have comparable performance based on the game benchmark suite used (66 combinations of games and resolutions).
Core Configuration
| GeForce GTX 1660 Ti | GeForce GTX 1070 | |||
|---|---|---|---|---|
| GPU Name | TU116 (TU116-400-A1) | vs | GP104 (GP104-200-A1) | |
| Fab Process | 12 nm | vs | 16 nm | |
| Die Size | 284 mm² | vs | 314 mm² | |
| Transistors | 6,600 million | vs | 7,200 million | |
| Shaders | 1536 | vs | 1920 | |
| Compute Units | 24 | vs | 15 | |
| Core clock | 1500 MHz | vs | 1506 MHz | |
| ROPs | 48 | vs | 64 | |
| TMUs | 96 | vs | 120 |
Memory Configuration
| GeForce GTX 1660 Ti | GeForce GTX 1070 | |||
|---|---|---|---|---|
| Memory Type | GDDR6 | vs | GDDR5 | |
| Bus Width | 192 bit | vs | 256 bit | |
| Memory Speed | 1500 MHz 12000 MHz effective |
vs | 2002 MHz 8008 MHz effective |
|
| Memory Size | 6144 Mb | vs | 8192 Mb |
Additional details
| GeForce GTX 1660 Ti | GeForce GTX 1070 | |||
|---|---|---|---|---|
| TDP | 120 watts | vs | 150 watts | |
| Release Date | 22 Feb 2019 | vs | 10 Jun 2016 |
-
GeForce GTX 1660 Ti
84.
96 GP/s -
GeForce GTX 1070
96.40 GP/s
GigaPixels — higher is better
-
GeForce GTX 1660 Ti
169.90 GT/s
-
GeForce GTX 1070
180.70 GT/s
GigaTexels — higher is better
-
GeForce GTX 1660 Ti
288.00 GB/s
-
GeForce GTX 1070
256.30 GB/s
GB/s — higher is better
-
GeForce GTX 1660 Ti
5437.00 GFLOPs
-
GeForce GTX 1070
5783.00 GFLOPs
GFLOPs — higher is better
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
44
-
GeForce GTX 1070
43
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
91
-
GeForce GTX 1070
86
FPS (higher is better)
Ultra quality TTA DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
90
-
GeForce GTX 1070
87
FPS (higher is better)
Ultra High Quality, TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
103
-
GeForce GTX 1070
104
FPS (higher is better)
Ultra High Quality TAA, Bahrain, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
93
-
GeForce GTX 1070
101
FPS (higher is better)
Windows 10 x64, Ultra quality, DirectX12
-
GeForce GTX 1660 Ti
40
-
GeForce GTX 1070
38
FPS (higher is better)
Very High Quality TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
104
-
GeForce GTX 1070
100
FPS (higher is better)
Highest quality DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
108
-
GeForce GTX 1070
107
FPS (higher is better)
Ultra quality DirectX12 Windows10 x64
-
GeForce GTX 1660 Ti
85
-
GeForce GTX 1070
86
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
71
-
GeForce GTX 1070
55
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
97
-
GeForce GTX 1070
93
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
68
-
GeForce GTX 1070
76
FPS (higher is better)
DX11, Max Details, 16:1 AF, 2xMSAA
-
GeForce GTX 1660 Ti
124
-
GeForce GTX 1070
152
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
105
-
GeForce GTX 1070
103
FPS (higher is better)
Ultra Detail,16:1 AF, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
47
-
GeForce GTX 1070
43
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
36
-
GeForce GTX 1070
32
FPS (higher is better)
Ultra Quality, HR Textures, DirectX11, Windows 10 x64
-
GeForce GTX 1660 Ti
80
-
GeForce GTX 1070
81
FPS (higher is better)
UltraTX Max Quality, Vulkan, Windows 10 x64
-
GeForce GTX 1660 Ti
46
-
GeForce GTX 1070
40
FPS (higher is better)
Max Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
98
-
GeForce GTX 1070
100
FPS (higher is better)
Highest detail, Pure Hair On, HBAO+, DirectX 12, Windows 11 x 64
-
GeForce GTX 1660 Ti
81
-
GeForce GTX 1070
74
FPS (higher is better)
Ultra Quality, DirectX12, Async Compute ,Windows 10×64
-
GeForce GTX 1660 Ti
102
-
GeForce GTX 1070
103
FPS (higher is better)
DX11,Max Details, 16:1 HQ-AF, +AA
-
GeForce GTX 1660 Ti
91
-
GeForce GTX 1070
83
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
34
-
GeForce GTX 1070
34
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
69
-
GeForce GTX 1070
67
FPS (higher is better)
Ultra quality TTA DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
69
-
GeForce GTX 1070
67
FPS (higher is better)
Ultra High Quality, TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
74
-
GeForce GTX 1070
77
FPS (higher is better)
Ultra High Quality TAA, Bahrain, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
71
-
GeForce GTX 1070
69
FPS (higher is better)
Windows 10 x64, Ultra quality, DirectX12
-
GeForce GTX 1660 Ti
26
-
GeForce GTX 1070
23
FPS (higher is better)
Very High Quality TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
71
-
GeForce GTX 1070
72
FPS (higher is better)
Highest quality DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
78
-
GeForce GTX 1070
77
FPS (higher is better)
Ultra quality DirectX12 Windows10 x64
-
GeForce GTX 1660 Ti
56
-
GeForce GTX 1070
56
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
52
-
GeForce GTX 1070
37
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
70
-
GeForce GTX 1070
68
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
48
-
GeForce GTX 1070
53
FPS (higher is better)
DX11, Max Details, 16:1 AF, 2xMSAA
-
GeForce GTX 1660 Ti
94
-
GeForce GTX 1070
101
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
68
-
GeForce GTX 1070
65
FPS (higher is better)
Ultra Detail,16:1 AF, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
35
-
GeForce GTX 1070
34
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
26
-
GeForce GTX 1070
21
FPS (higher is better)
Ultra Quality, HR Textures, DirectX11, Windows 10 x64
-
GeForce GTX 1660 Ti
56
-
GeForce GTX 1070
55
FPS (higher is better)
UltraTX Max Quality, Vulkan, Windows 10 x64
-
GeForce GTX 1660 Ti
35
-
GeForce GTX 1070
32
FPS (higher is better)
Max Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
65
-
GeForce GTX 1070
60
FPS (higher is better)
Highest detail, Pure Hair On, HBAO+, DirectX 12, Windows 11 x 64
-
GeForce GTX 1660 Ti
52
-
GeForce GTX 1070
49
FPS (higher is better)
Ultra Quality, DirectX12, Async Compute ,Windows 10×64
-
GeForce GTX 1660 Ti
73
-
GeForce GTX 1070
73
FPS (higher is better)
DX11,Max Details, 16:1 HQ-AF, +AA
-
GeForce GTX 1660 Ti
64
-
GeForce GTX 1070
65
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
12
-
GeForce GTX 1070
21
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
37
-
GeForce GTX 1070
37
FPS (higher is better)
Ultra quality TTA DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
40
-
GeForce GTX 1070
40
FPS (higher is better)
Ultra High Quality, TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
41
-
GeForce GTX 1070
43
FPS (higher is better)
Ultra High Quality TAA, Bahrain, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
40
-
GeForce GTX 1070
46
FPS (higher is better)
Windows 10 x64, Ultra quality, DirectX12
-
GeForce GTX 1660 Ti
10
-
GeForce GTX 1070
10
FPS (higher is better)
Very High Quality TAA, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
38
-
GeForce GTX 1070
39
FPS (higher is better)
Highest quality DirectX11 Windows10 x64
-
GeForce GTX 1660 Ti
43
-
GeForce GTX 1070
44
FPS (higher is better)
Ultra quality DirectX12 Windows10 x64
-
GeForce GTX 1660 Ti
28
-
GeForce GTX 1070
28
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
25
-
GeForce GTX 1070
26
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
37
-
GeForce GTX 1070
37
FPS (higher is better)
Ultra Quality, DirectX12, Windows 10 x64
-
GeForce GTX 1660 Ti
26
-
GeForce GTX 1070
29
FPS (higher is better)
DX11, Max Details, 16:1 AF, 2xMSAA
-
GeForce GTX 1660 Ti
47
-
GeForce GTX 1070
48
FPS (higher is better)
Ultra quality — DirectX12 Windows 11
-
GeForce GTX 1660 Ti
35
-
GeForce GTX 1070
35
FPS (higher is better)
Ultra Detail,16:1 AF, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
22
-
GeForce GTX 1070
21
FPS (higher is better)
Ultra Quality, DirectX11, Windows 10×64
-
GeForce GTX 1660 Ti
13
-
GeForce GTX 1070
12
FPS (higher is better)
Ultra Quality, HR Textures, DirectX11, Windows 10 x64
-
GeForce GTX 1660 Ti
33
-
GeForce GTX 1070
32
FPS (higher is better)
UltraTX Max Quality, Vulkan, Windows 10 x64
-
GeForce GTX 1660 Ti
24
-
GeForce GTX 1070
19
FPS (higher is better)
Max Quality, DirectX12, Windows 10×64
-
GeForce GTX 1660 Ti
32
-
GeForce GTX 1070
33
FPS (higher is better)
Highest detail, Pure Hair On, HBAO+, DirectX 12, Windows 11 x 64
-
GeForce GTX 1660 Ti
26
-
GeForce GTX 1070
25
FPS (higher is better)
Ultra Quality, DirectX12, Async Compute ,Windows 10×64
-
GeForce GTX 1660 Ti
41
-
GeForce GTX 1070
41
FPS (higher is better)
DX11,Max Details, 16:1 HQ-AF, +AA
-
GeForce GTX 1660 Ti
34
-
GeForce GTX 1070
39
FPS (higher is better)
| VS | ||
| GeForce GTX 1660 Ti | Radeon RX 5500 XT |
| VS | ||
| GeForce GTX 1660 Ti | GeForce GTX 1660 SUPER |
| VS | ||
| GeForce GTX 1070 | Radeon RX 5500 XT |
| VS | ||
| GeForce GTX 1070 | GeForce GTX 1660 SUPER |
| VS | ||
| Radeon RX 6600 | Radeon RX 6600 XT |
| VS | ||
| Radeon RX 6600 XT | GeForce RTX 2060 |
Please enable JavaScript to view the comments powered by Disqus.
newer GPU or more VRAM? GECID.com.
::>Video cards
>2023
> GeForce GTX 1660 SUPER vs. GeForce GTX 1070 2023: newer GPU or more VRAM?
03-03-2023
We continue to deal with popular graphics accelerators according to Steam. In the last review, we already checked if the GeForce RTX 2060 is up to date? Today we offer to consider another popular solution — the GeForce GTX 1660 SUPER. And to make it more interesting, let’s compare it with the GeForce GTX 1070, because on the used market these solutions cost almost the same: one is newer, and the other has more video memory. What’s better?
Let’s start with the GeForce GTX 1660 SUPER. This card is built on the same architecture as the older RTX 2000 line, but with RT blocks turned off.
The TU116 chip is installed on board. It received 1408 CUDA cores, 88 texture cores, 48 raster cores. All this beauty is seasoned with 6 GB of GDDR6 video memory connected via a 192-bit data bus with a bandwidth of 336 GB/s.
In turn, her opponent belongs to the already legendary 1000 series. The GeForce GTX 1070 of the previous Pascal architecture is equipped with a GP104 chip. On board is a team of 1920 CUDA cores, 120 texture and 64 raster units. The board also houses 8 GB of GDDR5 memory, old by today’s standards, running on a 256-bit bus with a bandwidth of 256 GB/s.
The GeForce GTX 1070 seems to have the upper hand in almost all key quantitative parameters, but our tests will show if this will lead to success compared to the more modern GeForce GTX 1660 SUPER in terms of architecture.
Let’s now look at the example of what we will test these cards with. The older solution is represented by the GeForce GTX 1070 DUAL version from Palit. For effective cooling of all components, it has a two-section cooling system with a massive plate in contact with three heat pipes and a main radiator. And so that the card does not suffocate, there are two massive proprietary fans.
This cooler keeps temperatures at 32°C idle and 73°C with a Hot Spot of 84°C under load.
Newer graphics adapter will be from the same vendor — Palit GeForce GTX 1660 SUPER with GAMINGPRO prefix. Its eponymous cooling system has visually changed in design, but the radiator has retained its dimensions. Unless they took one tube and increased the contact plate with the GPU, memory and VRM. But what remains unchanged is two fans for blowing the radiator. These simplifications did not affect the temperature much: 33°C at idle and 65-78°C under load is a decent result.
TeraFlops provided us with these video cards for testing. This is one of the few stores in Ukraine that gives a 12-month warranty for remanufactured components and 3 months for used ones. Also here you can find ready-made PCs at reasonable prices. And if you just like to look at interesting assemblies and components, subscribe to the store’s Instagram page.
Test bench
Before proceeding to compare the selected cards, let’s recall the test bench configuration.
Based on the ASUS ROG Strix B550-F Gaming WI-FI motherboard.
In order for both opponents to reveal their best qualities, one of the most popular mainstream Ryzen 5 5600X processors was chosen.
Added a Viper Steel 16 GB memory kit with two sticks at a frequency of 3600 MHz.
We used a be quiet! SHADOW ROCK TF 2.
The system, tests and games were installed on a 2 TB M.2 P300 series drive from PATRIOT.
The Seasonic PRIME TX-750 provided 750W of power, which is more than enough for this configuration.
The Chieftec Libra LF-02B has become a cozy home for components. You can find a detailed review of it on the website.
Chieftec AF-1125PWM fans were added to keep the case fresh.
Synthetic
Let’s go over synthetics very briefly. Only the Port Royal test stands out from the general picture, since it is run on such cards with software trace emulation.
So, here the GTX 1660 SUPER outpaced the GTX 1070 by 16%. In other tests, the difference was not significant. On average, we fix an insignificant, but still 5% advantage in favor of the GTX 1660 SUPER.
Games
Now let’s move on to games. The cards were tested in Full HD resolution. Where possible, the entire available video buffer was used.
So in Apex Legends at maximum graphics settings, both opponents showed brilliant gameplay. And only occasionally slipped minor twitches. The only thing in which the GTX 1070 outperformed the GTX 1660 SUPER was in the 0.1% Low column by a significant 30%.
And if you use all 8 GB of video memory, then a more experienced card turned out to be better by 2-33%.
Going ultra in Battlefield 2042 . There, the GTX 1660 SUPER was able to cross the 60 fps border. This made it possible to get a 3-4% advantage over the old GTX 1070, but the latter’s video buffer helped it outperform its rival by 2% in drawdowns.
What about Batla’s main competitor? With all the same settings, the player in Call Of Duty: Warzone 2.0 felt pretty average. We were constantly pursued, then subsidence to 45 FPS, then light freezes. Despite this, the difference in our confrontation reached 4-19% on the main indicators in favor of the GTX 1660 SUPER.
In Fortnite on epic graphics settings, both opponents were equal in drawdowns. And according to other indicators of the counter, the GTX 1660 SUPER turned out to be better by 4-9%. However, in the case of the GTX 1070, nothing prevents you from playing with a sufficient level of comfort.
But where the more experienced card was able to get ahead of its rival, it was in Overwatch 2 with epic settings. Here, her 8 GB video buffer was not needed. But what DX11 liked was the greater number of blocks in the GPU, which gave a 10-12% advantage.
In A Plague Tale: Requiem, had to go down to medium settings to get the cards closer to today’s desired 60 fps.
But, despite a temporary dip around 45 FPS, the gameplay was more or less stable. The difference between competitors was 2-9% in favor of the GTX 1660 SUPER.
Moving on to the noisy world of Cyberpunk 2077 . The picture on medium settings is a little better than in the previous game, but in general it is very similar. However, this only applies to the GTX 1660 SUPER, it more readily draws the streets of Night City, with a 22% lead over the opponent. For the GTX 1070, I want to lower the settings or enable FSR.
In Dying Light 2 , the newer architecture of the GTX 1660 SUPER gave it a 5-8 percent advantage across the board. Sometimes the 8 GB of VRAM in the GTX 1070 made itself felt and the gameplay with its participation kept within 57 fps.
Drunk Grandfather remake is too heavy for these video cards. Neither the architectural solutions of the GTX 1660 SUPER, nor the larger amount of VRAM in the GTX 1070 saved them from subsidence to 15-17 fps in 0.
1% Low. It was very difficult for the senior adapter. Interestingly, lowering the settings to low added only 2-3 FPS, which did not affect the situation in any way.
On the Forza Horizon 5 racing tracks with ultra settings, both cards went almost wheel to wheel. But in difficult places, the GTX 1070 sagged about 10 fps more noticeably, which eventually allowed the GTX 1660 SUPER to outperform its rival by 11-13%.
But the next race in Need For Speed: Unbound with ultra graphics was a pleasant surprise. Both contestants showed similar results. The difference, if it was, is not significant; without monitoring, we would not have noticed it.
In the rather controversial new Forspoken , which mimics the magic from the Harry Potter universe, our cards held at 50-57 fps on average. But that’s just low settings! Maybe FSR needed to be added? An interesting point — on the GTX 1070, most of the textures loaded in the best quality.
This can be clearly seen from the amount of video memory used, although the settings are the same.
In the fifth part of the story of the Ghost of Sparta, that is God of War with high settings, our opposition can be said to come to naught. A meager difference of 1-2 frames, or insignificant 2-4%, balances somewhere at the level of statistical error.
Also, our competitors passed the test from the trainee «friendly neighbor» in the game Marvel’s Spider-Man: Miles Morales . I was pleasantly surprised by the ability to choose even very high graphics settings. The GTX 1660 SUPER was 8-13% faster on the streets of New York than the GTX 1070. However, the latter also has a good result.
And finally, let’s take a walk through the streets and fields of sunny Tussaint in the «next-gen» interpretation of the third Witcher. At medium graphics settings, because with the highest both contestants it becomes quite difficult.
As a result, we have almost the same values in the 0.1% Low column and only a 3% advantage in favor of the newer GPU in terms of average.
Results
Shall we summarize? It is quite expected that both in synthetics and in games we got similar results. Despite the smaller video buffer, the newer architecture still allowed the GeForce GTX 1660 SUPER to keep its lead. We are talking about an average 7% advantage over the GeForce GTX 1070. Although, to the credit of the latter, she fought to the end. Despite this, both opponents are still viable options for thrifty gamers as they will allow you to run all popular games.
“How relevant in terms of money?” — you ask. Judge for yourself. Both cards can be found on the used market today with a price tag of 6,000 UAH. Although for the GeForce GTX 1660 SUPER they still ask a little more. Because it’s average and slightly more productive (keep in mind a fresher GPU and faster VRAM). Although the experienced GTX 1070 looks better in terms of quantitative characteristics, it even has a larger video buffer.
Therefore, the final choice will be influenced not only by the price at the time of purchase, but also by the condition of the video card itself — if possible, we inspect the desired instance before buying and agree on a check. Second-hand is Second-hand.
But buying a new GeForce GTX 1660 SUPER, for which they sometimes ask for more than 11,000 UAH, does not look like a very rational idea. In this case, find an opportunity to save up for a GeForce RTX 2060 or GeForce RTX 3050, or Radeon RX 6600. These will be more productive and more technologically advanced options! Although, again, everything will depend on price and availability, as well as promotions and discounts.
Author: Yuri Shcherbak
Also suggest reading:
Radeon RX 6500 XT vs GeForce GTX 1650 vs GeForce GTX 1050 Ti test & comparison | PCIE 4.
0 vs 3.0: maybe everyone was wrong?
GeForce RTX 3080 12GB vs. GeForce RTX 3080 10GB, GeForce RTX 3080 Ti, and Radeon RX 6900 XT
GeForce RTX 3050 in PCIe 3.0 vs. PCIe 4.0 vs. RTX 3060, RTX 2060, GTX 1660 Ti, and RX 6600
Comparison NVIDIA GeForce GTX 1660 Ti Mobile vs NVIDIA GeForce GTX 1070 Ti which is better?
%
DeviceList score
vs
NVIDIA GeForce GTX 1070 Ti
87%
DeviceList score
We compared NVIDIA GeForce GTX 1660 Ti (mobile) and NVIDIA GeForce GTX 1070 Ti and put together a list of benefits and a comparison table for you. Find out which one to choose in 2023.
Benefits NVIDIA GeForce GTX 1660 Ti Mobile
|
Value for money |
|
60.4%
14.2% (30.7%) better than vs 46.2% |
|
Release price |
|
$229
-170 $ (-42.6%) better than vs 399 $ |
|
Process |
|
12 nm
-4 nm (-25%) better than vs 16 nm |
|
Memory frequency |
|
12000 MHz
At 3992 MHz (49. vs 8008 MHz |
|
Memory bandwidth |
|
288
31.7 (12.4%) better than vs 256.3 |
9%) better than Benefits NVIDIA GeForce GTX 1070 Ti
Comparison winner
|
Number of shaders |
|
2432
896 (58.3%) better than vs 1536 |
|
Core frequency |
|
1607 MHz
152 MHz (10.4%) better than vs 1455 MHz |
|
Boost frequency |
|
1683 MHz
93 MHz (5.8%) better than vs 1590 MHz |
|
Maximum memory |
|
8 GB
2 GB (33.3%) better than vs 6 GB |
|
Memory bus width |
|
256 bit
64 bit (33.3%) better than vs 192 bit |
| General information | |
|
Price-quality ratio The sum of all the advantages of the device divided by its price. |
|
| 60.4%
14.2% (30.7%) better than |
46.2% |
|
Architecture |
|
| Turing | Pascal |
|
Codename |
|
| TU116 | GP104 |
|
Type |
|
| Notebook | Desktop |
|
Release price |
|
| $229
-170 $ (-42.6%) better than |
$399 |
|
Number of shaders |
|
| 1536 | 2432
896 (58.3%) better than |
|
Core clock |
|
| 1455MHz | 1607 MHz
152 MHz (10.4%) better than |
|
Boost frequency |
|
| 1590 MHz | 1683 MHz
93 MHz (5. |
|
Number of transistors |
|
| 6.600 million | 7.200 million |
|
Process |
|
| 12 nm
-4 nm (-25%) better than |
16 nm |
|
Interface |
|
| PCIe 3.0 x16 | PCIe 3.0 x16 |
|
Power Demand (TDP) The calculated heat output shows the average heat dissipation in load operation, |
|
| n/a | 180 W |
|
Length |
|
| n/a | 267 mm |
|
Additional power connectors |
|
| n/a | 1x 8-pin |
|
G-SYNC Ready NVIDIA G-SYNC technology delivers a smooth gaming experience with variable refresh rates and the elimination of visual artifacts. |
|
|
Multi Monitor |
|
| n/a | + |
|
VR Ready Technology from NVIDIA that gives manufacturers access to Multi res Shading, Context Priority, and GPU Direct virtual reality technologies. |
|
| n/a | + |
|
Vulkan NVIDIA’s Vulkan technology allows developers to gain low-level access to the GPU to optimize graphics commands (better than OpenGL and Direct3D APIs). |
|
| n/a | + |
|
Multi-Projection NVIDIA’s Multi-Projection technology improves graphics performance when building virtual worlds. Images for the left and right eyes are calculated simultaneously. |
|
| n/a | + |
|
Video connectors |
|
| No outputs | 1x DVI, 1x HDMI, 3x DisplayPort |
|
DirectX |
|
| 12 (12_1) | 12 (12_1) |
|
Floating point performance |
|
| n/a | 8. 186 gflops |
The more%, the better the quality per unit price in comparison with all analogues. 
8%) better than 
| Benchmarks |
| Memory
Memory type |
|
| GDDR6 | GDDR5 |
|
Maximum memory Large video memory allows you to run demanding games with lots of textures, |
|
| 6 GB | 8 GB
2 GB (33.3%) better than |
|
Memory bus width The larger the video memory bus width, the more data is transferred to the GPU per unit of time and the better performance in demanding games. |
|
| 192 bit | 256 bit
64 bit (33.3%) better than |
|
Shared memory |
|
| — | — |
|
Memory frequency A high memory frequency has a positive effect on the speed of a video card with a large amount of data. | |
