Performance Analysis

The RTX 2060 handles all 1080p gameplay with ease, while at 1440p things are tougher but still managed very well with consistently smooth frame rates across the board, excluding Total War: Warhammer II where driver issues still seem to plague Nvidia’s benchmark results. A 4K resolution, however, is too much for the RTX 2060, at least when using Ultra settings like we do and like most prefer to. The option to lower settings is always there, of course, but we reckon most will be targetting high frame rate 1080p or 1440p gaming with RTX 2060, and rightly so.

The Palit RTX 2070 Dual serves as our reference model, since the RTX 2070 Founders Edition is absent from our roster. Compared to this card, Palit’s offers 15 percent more performance on average yet is 39 percent more expensive, so the performance gap is actually closer than we thought it would be. The RTX 2070 does extend its average uplift slightly as you go up the resolution scale, giving some credence to the idea that 6GB is a bit limiting and could be more so in the future, but we’re only talking a few percentage points, and there is no clear drop-off for this card even at 4K in any test, so it's equally important not to over-inflate this finding.

Versus previous-generation Nvidia hardware, the RTX 2060 FE has the GTX 1070 Ti FE beat in nearly every test and is faster by seven percent on average, but a GTX 1080 OC couldn’t be caught except in certain titles that favour RTX hardware like Wolfenstein and VRMark Cyan Room.  Meanwhile, RTX 2060 offers 59 percent more performance than its predecessor by name, the GTX 1060 6GB.

Next to rival AMD, the RTX 2060 sits between RX 590 and Vega 56 in price (albeit much closer to the latter) but tends to be in between the two Vega parts when it comes to performance. For instance, it is 42 percent quicker than the RX 590 on average, and interestingly its lead is actually a bit higher at 1440p than at 1080p despite a 2GB advantage in memory size for AMD here. It isn’t a straight victory over its closest rival, Vega 56, meanwhile (see Deus Ex and Warhammer II, for example), but by and large this is the faster card, with our results averaging out to an eight percent lead for RTX 2060. Vega 64 at least manages to keep ahead most of the time, but even then RTX 2060 is never far behind.

Boosting under sustained load saw the card reach speeds usually between 1,845MHz and 1,875MHz, with power most often listed as the performance cap. This is roughly 100MHz faster than we saw from Palit’s reference RTX 2070, which helps explain why the gap is smaller than it appears on paper.

While beating the GTX 1070 Ti on performance overall, power consumption ended up being a little lower – can’t argue with that. As usual, Nvidia wipes the floor with AMD cards here, with even the RX 590 sending our system consumption to 315W compared to 259W with this part.

A delta temperature value of 48°C is good but nothing amazing. All three RTX Founders Edition cards are within 4°C of one another, suggesting Nvidia is targetting a similar temperature each time. With its relatively low power requirements, the RTX 2060 only needed its onboard fans to spin up to around 1,600 RPM and stayed pleasingly quiet throughout testing. Idle, you can expect a speed of 1,200 RPM. It’s a shame Nvidia isn’t able to deploy semi-passive cooling, but the card is nonetheless very quiet indeed when idle.

Overclocked by 125MHz, our card was boosting to around 1,950MHz depending on the game – we didn’t see the card spending much time over 2GHz. Fan speeds and noise didn’t vary all that much, and neither did power consumption, so there isn’t much cost to overclocking this card. Performance went up by between four and eight percent, which isn’t bad and isn’t amazing, but it does eat away even more at the leads held by Vega 64 and RTX 2070, even managing to surpass the former in Battlefield 1.