The rumours were true: the GeForce RTX 3060 Ti is a real graphics card, and we’ve been testing it for the past week. Nvidia promises RTX 2080 Super level performance in a smaller, cheaper and more efficient Ampere design, and – spoiler alert – that’s exactly what our testing shows they’ve delivered.
For £370 in the UK, $400 in the US and around €399 in Europe, the 3060 Ti isn’t quite in mainstream territory, but it’s significantly faster than the RTX 2060 Super it replaces at the same price-point and AMD’s competitive Big Navi architecture has yet to debut at anything below the $580/£530 price point of the RX 6800. If the Green Team is able to actually produce these cards in volume – and sell them to actual customers rather than bot farms – they could be onto a winner.
In terms of the specs and underlying architecture, the RTX 3060 Ti uses the same GA104 GPU as the RTX 3070, but with fewer CUDA cores – 4864 versus 5888. The card also operates at slightly slower clock speeds (1665MHz boost versus 1725MHz) to fit into a 20W lower TDP (200W vs 220W). The memory subsystems are unchanged however, with both cards sporting the same 8GB GDDR6 operating at 448GB/s. It’s good to see 8GB of VRAM becoming the new standard, with all next-gen cards from both teams providing at least that much thus far.
The comparison against older generation cards is more interesting. The 3060 Ti manages to more than double the number of graphics cores of the 2060 Super, in a smaller die that consumes only a tad more power. Looking further back, at the GTX 1060, and the multiplier is closer to 4x – with a corresponding increase to transistor count, courtesy of the shift from the 16nm with Pascal to 12nm with Turing and now 8nm with Ampere.
As well as improvements to compute performance, all of the usual features of Nvidia’s Ampere architecture are present and correct, including next-generation ray tracing and tensor cores, so we can expect more noticeable performance uplifts in RT and AI accelerated workloads. We’ve covered Ampere in more detail in our earlier RTX 3070, 3080 and 3090 reviews, so let’s move onto the card’s physical design.
As usual, we’re testing the Founders Edition of the card, which comes with the same gorgeous industrial design and in the same compact dimensions as the RTX 3070. We have two axial fans in a ‘flow-through’ configuration, with the small, pennant-shaped motherboard design and miniature 12-pin power connector allowing the final third of the card to be wholly used for cooling. Even the I/O is arranged to maximise airflow, with a single row of display outputs (three DisplayPort 1.4, one HDMI 2.1) sitting beneath a 16×5 grid of ventilation cutouts.
As with other RTX 30-series cards, the 3060 Ti is a PCIe 4.0 device but works in PCIe 3.0 motherboards without any loss in performance in the vast majority of gaming workloads. Another invisible feature is the AV1 decode support, which should allow sites like Twitch and Netflix to up their resolution, frame-rate and bitrate substantially at a given bandwidth – or decrease bandwidth by up to 50 per cent while keeping these metrics the same. AV1 support isn’t essential now, and is available on both AMD and Nvidia’s new cards, but it could be a nice cherry on top for anyone choosing to upgrade once it hits the mainstream.
One of the biggest questions we had when testing the 3060 Ti – beyond its gaming performance, which we’ll get to soon – surrounded its power efficiency. The RTX 3070 was substantially more efficient than its more powerful siblings, so is the 3060 Ti more efficient still?
To answer this question, we use Nvidia’s Power Capture Analysis Tool, or PCAT. This is an interposer board that sits between the PCIe slot and the graphics card, as well as between the supplementary 8-pin power input used by the 3060 Ti and our power supply. This way, we can measure the number of watts drawn by the card itself, rather than the load of the full system which can vary naturally over time, and plot this power draw precisely against frame-rates to get a sense of how much power is being used to create each frame.
In Death Stranding, the RTX 3060 Ti almost equals the RTX 3070, requiring 3.064 joules per frame to render our test scene compared to 2.915 joules per frame for the 3070. AMD’s Big Navi graphics cards perform better in Death Stranding than Nvidia’s Ampere, so we see an even better efficiency rating for the RX 6800 and RX 6800 XT. So the 3060 Ti does use less power than the 3070, but the decrease in performance is larger than the decrease in power usage – at least in this test.
Gears 5 is more hopeful, with the 3060 Ti setting a new record here of 3.163 joules per frame, with the 3070 requiring 14 per cent more power per frame and the RX 6800 requiring around 20 per cent more. Normally we’d expect cards in the same family to be arranged in the same order in each game, so it’s unusual to see the 3060 Ti lead against the 3070 in one test and trail in another. However, retests bore out our initial results, allowing the 3060 Ti to claim the ‘most efficient Ampere GPU’ title in at least one game. It’s worth looking back at the RTX 2070 FE as well, which requires 50 per cent more joules for each frame – Ampere and its shift to 8nm has really had solid results in terms of power efficiency.
With our brief power testing concluded, it’s time to confirm the specifications of our test rig. You’ll see a few new components here if you’ve not checked out our RTX 30-series reviews yet, but for those of you that are familiar there are no surprises here. We have a Core i9 10900K is locked to a 5GHz all-core frequency and cooled by a 240mm Alpacool Eisbaer Aurora AiO – which keeps the overclocked system at around 75C under full load.
The 10900K is backed by an Asus Maximus 12 Extreme Z490 motherboard and two 8GB sticks of G.Skill Trident Z Royal 3600MHz CL16. Our games are run from a capacious 2TB Samsung 970 Evo Plus NVMe drive provided by Box. The whole rig is powered by a 850W gold-rated Gamer Storm power supply.
With the stage set, let’s begin the show with some gaming benchmarks.