Faster than the 2080 Super and easily outperforms 1080Ti •

The rumors were true. The GeForce RTX3060Ti is a real graphics card and we’ve been testing it for the past week. Nvidia promises RTX 2080 super-level performance with a smaller, cheaper and more efficient Ampere design. And spoiler alerts are exactly what our tests have provided.

£ 370 in the UK, $ 400 in the US, € 399 in Europe, and the 3060Ti is not. Pretty Although in the mainstream territory, it is significantly faster than the RTX 2060 Super and can be replaced at the same price. AMD’s competitive Big Navi architecture has not yet debuted at the RX6800 for less than $ 580 / £ 530. In the case of a green team, these cards can actually be mass-produced and sold to real customers instead of bot farms, which could be a winner.

In terms of specifications and underlying architecture, the RTX 3060 Ti uses the same GA104 GPU as the RTX 3070, but with fewer CUDA cores-4864 vs. 5888. The card also works at slightly slower clock speeds (1665MHz boost vs. 1725MHz). 20W lower TDP (200W vs 220W). However, the memory subsystem has not changed and both cards are running at 448GB / sec with the same 8GB GDDR6. It’s nice to see 8GB of VRAM becoming the new standard and all next generation cards from both teams are offering that much, at least so far.

This content is hosted on an external platform and will only be displayed if you accept targeting cookies. Please enable cookies to view them.

RTX 3080 RTX 3070 RTX 3060 Ti RTX2060 Super RTX 2060 GTX 1060
GPU GA102 GA104 GA104 TU106 TU106 GP106
CUDA core 8704 5888 4864 2176 1920 1280
Memory bus 320 bits 256 bits 256 bits 256 bits 192 bits 192 bits
Bandwidth 760GB / sec 448GB / sec 448GB / sec 448GB / sec 336GB / sec 192GB / sec
Base clock 1440MHz 1500MHz 1410MHz 1470MHz 1365MHz 1506MHz
Boost clock 1710MHz 1725MHz 1665MHz 1650MHz 1680MHz 1709MHz
TDP 320W 220W 200W 175W 160W 120W
Die size 628mm2 392mm2 392mm2 445mm2 445mm2 200mm2
Transistor 28B 17.4B 17.4B 10.8B 10.8B 4.4B

The comparison with older generation cards is more interesting. The 3060 Ti is a small die that consumes slightly more power and can more than double the number of graphics cores in the 2060 Super. Looking back further, looking at the GTX 1060, the multiplier is nearly four times higher, and the number of transistors is increasing accordingly. This is due to the shift from Pascal’s 16nm to Turing’s 12nm and amperes to 8nm.

In addition to improving compute performance, all the usual features of Nvidia’s Ampere architecture, such as next-generation ray tracing and tensor cores, are present and correct, so they perform better with RT and AI accelerated workloads. Can be expected to improve. Now that we’ve covered Ampere in detail in previous reviews of the RTX 3070, 3080, and 3090, let’s move on to the physical design of the card.

As always, we’re testing the Founders Edition of a card with the same luxurious industrial design and the same compact dimensions as the RTX 3070. It has two axial fans in a “flow-through” configuration, a small pennant motherboard design and a miniature 12-pin power connector that allows the last third of the card to be used for complete cooling. Even the I / O are arranged to maximize ventilation, with a single row of display outputs (3 DisplayPort 1.4, 1 HDMI 2.1) placed under a 16×5 grid ventilation cutout.

Like other RTX 30 series cards, the 3060Ti is a PCIe 4.0 device, but it runs on PCIe 3.0 motherboards for most of the gaming workload without compromising performance. Another invisible feature is support for AV1 decoding. This allows sites such as Twitch and Netflix to significantly increase resolution, frame rate, and bitrate on certain bandwidths. Alternatively, you can reduce bandwidth by up to 50% while keeping these metrics the same. AV1 support isn’t currently required and is available on both AMD and Nvidia’s new cards, but it could be the best cherry for anyone who chooses to upgrade once it’s mainstream.

One of the biggest questions we had when testing the 3060 Ti was, as you can see, power efficiency, beyond game performance. The RTX 3070 was significantly more efficient than the more powerful siblings, but is the 3060 Ti even more efficient?

To answer this question, use Nvidia’s Power Capture Analysis Tool (PCAT). This is an interposer board located between the PCIe slot and the graphics card, and between the auxiliary 8-pin power input and power supply used on the 3060Ti. In this way, we measure the wattage consumed by the card itself, rather than the overall system load, which can change naturally over time, and plot this power consumption accurately against the frame rate. You can grasp the amount of power. It is used to create each frame.

Nvidia’s PCAT system uses a PCIe interface and power cable interposer board to measure the actual juice consumed by the GPU. Note that we were using the RTX3080PSU adapter here. The 3060Ti requires only one 8-pin PCIe power input and uses a different dongle.
Jules per frame RX 6800 XT RX 6800 RTX 3080 FE RTX 3070 FE RTX 3060 Ti FE RTX 2070 FE RX 5700 XT
Death Stranding 2.933 2.644 3.349 2.915 3.064 3.651 3.846
Percentage difference 110.9% 100% 126.6% 110.2% 115.9% 138.1% 145.5%
Jules per frame RX 6800 XT RX 6800 RTX 3080 FE RTX 3070 FE RTX 3060 Ti FE RTX 2070 FE RX 5700 XT
Gears 5 4.384 3.792 4.156 3.603 3.163 4.734 5.403
Percentage difference 138.6% 119.9% 131.4% 113.9% 100% 149.7% 170.82%

In Death Stranding, the RTX3060Ti is about the same as the RTX3070, which requires 3.064 joules per frame to render the test scene, compared to 2.915 joules per frame for the 3070. AMD’s Big Navi graphics card outperforms Nvidia’s Ampere in Death Stranding. See even better efficiency evaluations for the RX6800 and RX6800XT. Therefore, the 3060 Ti uses less power than the 3070, but the performance degradation is greater than the power usage degradation, at least in this test.

Gears 5 is more hopeful, with the 3060 Ti setting a new record of 3.163 joules per frame, the 3070 requiring 14% power per frame and the RX 6800 requiring about 20% power. It’s unusual for a 3060 Ti to lead 3070 in one test and trail in another, as cards from the same family are usually expected to be placed in the same order in each game. However, the retest lost the first result, allowing the 3060Ti to win the title of “Most Efficient Ampere GPU” in at least one game. The RTX 2070 FE is also worth looking back. This requires 50% more joules per frame. Ampere and its shift to 8nm have had solid results in terms of power efficiency.

After a brief power test, check the test rig specifications. If you haven’t checked the RTX 30 series reviews yet, you’ll see some new components here, but for those familiar with it, it’s not surprising here. The Core i9 10900K is locked to all 5GHz core frequencies and is cooled by the 240mm Alpacool Eisbaer Aurora AiO. This keeps the overclocked system at about 75C at full load.

The 10900K is supported by an Asus Maximus 12 Extreme Z490 motherboard and two 8GB sticks, the G.Skill Trident Z Royal 3600MHz CL16. Our game runs from the large capacity 2TB Samsung 970 Evo Plus NVMe drive provided by Box. The entire rig is powered by a 850W gold rated gamerstorm power supply.

Once the stage is set, let’s start the show with some game benchmarks.

Nvidia GeForce RTX 3060 Ti Analysis

  • First, hardware and power analysis [This Page]

  • Doom Eternal, Control, Borderlands 3, Shadow of the Tomb Raider-Game Benchmark Part 1

  • Death Stranding, Far Cry 5, Hitman 2, Assassin’s Creed Odyssey-Game Benchmark Part 2

  • Metro Exodus, Dirt Rally 2, Assassin’s Creed Unity-Game Benchmark Part 3

  • Control, Metro Exodus, Battlefield 5-RTX Game Benchmark

  • Nvidia GeForce RTX 3060 Ti-Digital Foundry Verdict

Back to top button