Nvidia GeForce RTX 2060 Super / RTX 2070 Super review: timely upgrades
Product refresh or preemptive strike?
This was supposed to be AMD's moment - the time when Navi finally arrived to make Team Red competitive again towards the higher end of the graphics hardware market. Just days before the launch of the Radeon RX 5700 and RX 5700 XT though, Nvidia fired a new salvo: three 'tweener' cards designed to take the wind out of AMD's sails by offering superior performance at a dangerously similar price. In fact, the firm was confident in its new line-up that it even pulled its embargo forward - so products that many see as a reaction to AMD's Navi will actually be reviewed ahead of their prospective competition.
As Nvidia teased back in May, three cards form the Super line-up: the RTX 2060 Super, RTX 2070 Super and RTX 2080 Super. The 2080 Super is still waiting in the wings, but we've tested the first two members of the family and as you'll see from the numbers, they're pretty impressive.
Compared to their vanilla counterparts, the Super series cards are improved in almost every way. The new GPUs offer a greater number of CUDA cores, higher base clocks and even improved hardware-accelerated ray tracing abilities. However, these higher-clocked components do draw more power; a little more in the case of the 2060 Super and a lot more for the 2070 Super. The RTX 2060 Super also sports 8GB of VRAM, compared to the 6GB of its predecessor - a good addition bearing in mind that ray tracing in particular can push memory hard. Altogether, it's an impressive package that perhaps suggests Nvidia built in some headroom with their first-generation Turing cards.
Final pricing has been confirmed: it's £379/$399/€419 for the RTX 2060 Super, which sees a price bump over the vanilla card and £479/$499/€529 for the RTX 2070 Super, which is what Nvidia are currently asking for the original 2070 Founders Edition card. When we consider AMD's pricing for its new Navi graphics cards - $379 for the Radeon 5700 and $449 for the Radeon 5700 XT - it's clear that both companies have a decent offering, but AMD won't be able to gazump the original RTX 2060 and RTX 2070 cards as planned. Based on AMD's own benchmarks up against the older RTX cards, we should expect both RTX 2060 Super and its 2070S equivalent to match or beat Navi based on rasterisation performance alone, before we factor in the attraction of features like ray tracing, variable rate shading and DLSS. As usual though, wait for reviews and benchmarks before drawing any firm conclusions on the upcoming Navi vs Turing dust-up.
Returning to the matter at hand, curiously, the RTX 2060 Super uses the TU106 processor found on the original RTX 2060 and RTX 2070. The RTX 2070 Super actually uses the TU104 core that forms the basis of the RTX 2080 instead, explaining its significantly higher TDP - and the six-pin plus eight-pin power inputs on the Founders Edition card we were sent for review. Meanwhile, the RTX 2060 Super Founders Edition requires only a single eight-pin input, situated at the back of the card. In terms of I/O, the the 2070 Super FE includes three DisplayPorts, USB-C and HDMI, while the 2060 Super FE swaps one of the DisplayPorts for DVI.
It's worth taking a look at the specs table below to see exactly what each card delivers compared to its predecessors at the hardware level. For example, the pared-down 192-bit memory bus in the original RTX 2060 has been returned to full 256-bit strength for the 2060 Super, unlocking a corresponding bandwidth increase to 448GB/s that equals the RTX 2070 and RTX 2070 Super.
Meanwhile, the RTX 2070 Super makes its biggest strides elsewhere, moving from 2304 to 2560 CUDA cores plus a healthy increase in boost clocks, from 1620MHz to 1770MHz - and we regularly saw the card hit 1935MHz in our testing. In terms of ray tracing performance, the RTX 2060 Super has the same 6GR/s rating as the original RTX 2070, while the 2070 Super's 7GR/s puts it just a step behind the RTX 2080.
| RTX 2060 | 2060 Super | RTX 2070 | 2070 Super | RTX 2080 | |
|---|---|---|---|---|---|
| CUDA cores | 1920 | 2176 | 2304 | 2560 | 2944 |
| Giga Rays/sec | 5 | 6 | 6 | 7 | 8 |
| VRAM | 6GB GDDR6 | 8GB GDDR6 | 8GB GDDR6 | 8GB GDDR6 | 8GB GDDR6 |
| Memory Bus | 192-bit | 256-bit | 256-bit | 256-bit | 256-bit |
| Bandwidth | 336GB/s | 448GB/s | 448GB/s | 448GB/s | 448GB/s |
| Boost Clock | 1680MHz | 1650MHz | 1620MHz | 1770MHz | 1710MHz |
| Processor | TU106 | TU106 | TU106 | TU104 | TU104 |
| TDP | 160W | 175W | 175W | 215W | 215W |
Now, let's take a closer look at the performance you can expect from these two cards in an array of the latest games. As well as the original RTX 2060 and RTX 2070, we'll compare the new Super cards to the best Pascal cards. Note that AMD's upcoming Navi offerings, the RX 5700 and RX 5700 XT, are not yet available and are therefore not included in our tests at present. However, we will redress this point once the 5700 series cards reach consumers; stay tuned. Important note: We've tested Founders Edition RTX cards across the board here, and while the original RTX releases featured small factory overclocks, the Super FE cards do not. This gives the older cards a very slight boost, but it doesn't seem to make too much difference to the general outlook.
Nvidia GeForce RTX 2060 Super / RTX 2070 Super Analysis
- Introduction, Hardware Breakdown [This Page]
- RTX vs RTX Super: Ray tracing performance analysis
- Assassin's Creed Odyssey/Unity, Battlefield 1 - Rasterisation Analysis Part 1
- Crysis 3, Far Cry 5, Ghost Recon Wildlands - Rasterisation Analysis Part 2
- Rise of the Tomb Raider, Shadow of the Tomb Raider, The Witcher 3 - Rasterisation Analysis Part 3
- Nvidia GeForce RTX 2060 Super/ RTX 2070 Super - the Digital Foundry verdict