Until September 2020, the RTX 2080 Ti was the fastest customer graphics card on the marketplace and retailed for $1000 – frequently more. Now, Nvidia’s brand-new RTX 3070 is here, with Team Green assuring the very same level of graphics efficiency at $500, half the cost. That’s a vibrant yet basic proposal that will no doubt stimulate a great deal of upgrades, so does it hold water? We’ve been evaluating the 3070’s video gaming efficiency and power performance for the previous week to discover.

For what it deserves, Nvidia definitely appears to be anticipating a great deal of need. The RTX 3070 was initially set to go reside in mid-October, however the launch was postponed by 2 weeks to develop stock of the brand-new cards and bulk up procedures versus automated purchases. The RTX 3080 and 3090 both offered out right away and stay near-impossible to discover 6 weeks post-launch, so let’s hope that schedule is enhanced this time around.

So presuming for a minute that cards are offered for purchase in any sort of sensible amount of time, how does the RTX 3070 provide the sort of efficiency gains Nvidia is declaring?

The response depends on architectural enhancements, primarily, consisting of a die avoid TSMC’s 14nm procedure to Samsung’s 8nm. That permits much more transistors to be loaded into a smaller sized area, opening both efficiency and performance gains. Combined with second-generation ray tracing and third-generation tensor cores, some seriously outstanding gen-on-gen enhancements can be attained – as you can see from our typically complimentary RTX 3080 and RTX 3090 evaluations.

The 3070 ought to press efficiency per dollar to even greater levels, while drawing less power (and for that reason creating less heat). It boasts 5344 CUDA cores compared to the 2304 in the RTX 2070, with approximately the very same clock speeds and a comparable GDDR6 memory user interface instead of the more pricey GDDR6X VRAM we saw on the 3080 and 3090. Compared to the RTX 2080 Ti, the 3070 has more CUDA cores however less power and less memory bandwidth, so it’ll be intriguing to see how it compares in various titles.

RTX 3080 RTX 3070 2080 Ti RTX 2080 RTX 2070 GTX 1070
GPU GA102 GA104 TU102 TU104 TU106 GP104
CUDA cores 8704 5888 4352 2944 2304 1920
VRAM 10GB GDDR6X 8GB GDDR6 11GB GDDR6 8GB GDDR6 8GB GDDR6 8GB GDDR5
Memory Bus 320-bit 256-bit 352-bit 256-bit 256-bit 256-bit
Bandwidth 760GB/s 448GB/s 616GB/s 448GB/s 448GB/s 256GB/s
Base Clock 1440MHz 1500MHz 1350MHz 1515MHz 1410MHz 1506MHz
Boost Clock 1710MHz 1725MHz 1545MHz 1800MHz 1620MHz 1683MHz
TDP 320W 220W 250W 225W 175W 150W
Die Size 628mm2 392mm2 754mm2 545mm2 445mm2 314mm2
Transistors 28B 17.4B 18.6B 13.6B 10.8B 7.2B

For our evaluation, we’re evaluating the RTX 3070 Founders Edition, which retails at the very same $500/£450 RRP as entry-level customized styles from Nvidia’s partners. That’s a modification from last-generation RTX 20-series FE designs, which included a factory overclock and a slightly greater power target however likewise cost around $100 more than the least expensive third-party equivalents.

Like Nvidia’s 3080 and 3090 FE designs, the 3070 Founders Edition utilizes a thick ‘flow-through’ cooling service, with a fairly little PCB and a great deal of metal fins. Unlike the greater designs though, the 3070 has both of its fans on the very same side and the card is visibly smaller sized – at around 24 by 11cm (approximately 9.5 by 4.5 inches), it’s basically the very same height and simply a centimetre longer than the outbound RTX 2070 FE card regardless of appealing substantially greater efficiency. It’s likewise substantially lighter than other Ampere cards, weighing simply 1035 grams compared to 1352 grams for the RTX 3080 FE and a monstrous 2157 grams for the 3090 FE.

In regards to I/O, we’re taking a look at the very same attire as the RTX 3080 and 3090: 3 DisplayPort 1.4 ports and one HDMI 2.1 port efficient in 4K 120Hz or 8K 60Hz with the best screen, however no USB-C VirtualLink (RIP). AV1 encoding is consisted of here too, which ought to permit websites like Netflix and Twitch to stream high-quality, high frame-rate video material to you while needing substantially lower bandwidth than present codecs. Finally, the card is PCIe 4.0 suitable however works simply great in PCIe 3.0 slots too so you aren’t anticipated to hurry out and purchase a brand-new motherboard.

The 3070 FE drinks power compared to its larger brother or sisters, needing just a single 8-pin power input – linked through the adapter in package to the very same 12-pin mini adapter as we saw on Nvidia’s RTX 3080 and 3090 cards. Those quicker designs needed 2 8-pin power inputs for their 12-pin adapters, so it’s not a surprise to discover that the 3070’s ranked TDP is 220W, substantially less than the 320W and 350W of the 3080 and 3090, respectively.

To put these power figures into point of view, we utilized Nvidia’s PCAT system, a board that sits in between the graphics card and its power inputs, to determine real power usage when video gaming. When twinned with in-game efficiency metrics, we can utilize this to exercise precisely the number of watts are needed to render each frame, and therefore get a step of real-world performance of the RTX 3070 and its rivals from Nvidia and AMD.

Watts Per Frame RTX 3090 FE RTX 3080 FE RTX 3070 FE RTX 2080 Ti FE RTX 2070 RX 5700XT
Death Stranding 3.209 3.349 2.915 3.562 3.651 3.846
Percentage Diff 110.1% 114.9% 100.0% 122.2% 125.2% 131.9%
Watts Per Frame RTX 3090 FE RTX 3080 FE RTX 3070 FE RTX 2080 Ti FE RTX 2070 RX 5700XT
Gears 5 4.055 4.156 3.603 4.266 4.734 5.403
Percentage Diff 112.5% 115.3% 100% 118.4% 131.4% 150.0%

The 3070 leaves to an early lead in a run of Gears 5, with the RTX 2080 Ti providing comparable efficiency however needing almost 20 percent more power to do so. The RTX 3080 and 3090 accomplish greater frame-rates however utilize a lot more power, offering the 3070 the performance edge. AMD’s newest high-end card at the time of this evaluation, the RX 5700 XT, needs 50 percent more power to render each frame.

We see a comparable lead to Death Stranding when it concerns the RTX 2080 Ti, which surpasses the RTX 3070 somewhat however draws more than 20 percent more power per frame. AMD cards are more effective in Death Stranding than Gears 5, however the RX 5700 XT still need 32 percent more watts per frame than the 3070. Of course, AMD’s brand-new Big Navi cards are possibly simply weeks away and ought to feature performance enhancements of their own. For the here and now however, the 3070 is our brand-new performance champ.

With the power screening out of the method, it’s time to move onto the video game criteria. You’ll see a couple of brand-new titles here if you have actually not had a look at our RTX 3080 and 3090 evaluations yet, along with a brand-new test platform.

While AMD’s Ryzen 3950X and its PCIe 4.0 compatibility were thought about for our graphics test bed, in the end we chose the Intel Core i9 10900K, which exceeded the 3950X in our video gaming criteria even with PCIe 4.0 cards like the RTX 30-series restricted to PCIe 3.0 speeds. At least in the meantime, raw CPU power appears to have a terrific impact on frame-rates than PCIe bandwidth – beyond a couple of particular situations.

The Core i9 10900K is locked to a 5GHz all-core frequency and backed by an Asus Maximus 12 Extreme Z490 motherboard and 2 8GB sticks of G.Skill Trident Z Royal 3600MHz CL16. Cooling responsibilities are dealt with by a 280mm AiO from Corsair. Our video games are ranged from a capacious 2TB Samsung 970 Evo Plus NVMe drive provided by Box. The entire rig is powered by a 1000W Corsair power supply.

Now that you understand topography, let’s enter some video gaming tests.

Nvidia GeForce RTX 3070 Analysis