Nvidia’s RTX 30-Series has been blowing our minds since its release, bringing out some of the best cards in a long time. We compare the RTX 3070 vs 3080.
Choosing the right graphics card for your system can be difficult because there is such a wide selection available.
The 30-Series marked a complete shift in architectural makeup, moving from the infamous Turing blueprint to the all-singing, all-dancing Ampere microarchitecture. A smaller 8-nm process node, 50% more CUDA cores, a third-generation Tensor core, and a second-generation RT core are just a few of the advances.
Based on the GA104-300-A1 variant of Ampere, the 3070 contains 5888 shading units, 184 texture units, 96 rendering units, and 184 Tensor cores.
The 3080 is the product of the GA102-200-KD-A1 Ampere variant containing a monstrous 8704 shader units – nearly 3000 more than the 3070. That’s a pretty big jump for a single GPU step in the same family. Additionally, it offers 88 texture units for high-resolution gaming, providing an additional layer of realism. Apart from 1MB of larger 5MB L2 cache and just shy of 100 more Tensor cores, these cards are otherwise quite similar.
The cooling process
Another huge improvement in Ampere is just how good the thermals of the 30-Series are. In addition to reaching 93°C before throttling, the 3070 is unlikely to exceed 80°C under load. If you pair it with a high-spec air cooler or liquid AIO, you can overclock by 5%, or that’s the performance of a 2080 Ti at a fraction of the cost.
Even though the 3080 draws 320 watts as opposed to the 3070’s 220-watt draw, it still maxes out around 79°C when overloaded at 93°C. You can achieve a +75MHz core overclock with a quality cooler, which results in slightly boosted frame rates and lower GPU temperatures.
The Galax Classic series includes a 3080 blower edition, while the Asus Turbo series offers a 3070 blower edition.
In the end, if these GPUs don’t fit into your case, they’re useless to you, so let’s take a look at the boring stuff. The RTX 3070 vs 3080 battle is tough, but there can be some significance in their sizes.
Despite being so powerful, the 3070 is quite modestly sized, making it a good choice for smaller ATX mid-tower cases. Measuring 9.5″ (L) x 4.4″ (W) x 2-slot, it easily replaces previous-generation graphics cards from mid- to high-tier.
With a size of 11.2″ x 4.4″ x 2-slot, the 3080 will need a large mid-tower case with strong brackets. For GPU sag prevention, you may want to invest in a support brace. The 3rd party iterations of the 3080 tend to be even larger, so make sure you take accurate case measurements.
FPS and resolution
So enough about physical attributes. Let’s get down to business… real-world gaming performance. The battle of the RTX 3070 vs 3080 aims to give you the best gaming experience possible.
If you’re a 1080p gamer, you’ll really need a monitor with a high refresh rate to get anywhere near the performance these cards are capable of. Having said that, we always advise buying the latest technology where budget allows to future-proof a potential monitor purchase in the future.
1440 x 900
You’re unlikely to find a game in which the 3070 doesn’t achieve the desired 60fps average. Even with RDRII, it averages 80fps, making it perfect for monitors with 60Hz 1440p resolutions.
RDRII has a frame rate that pushes beyond 100fps, and Shadow of the Tomb Raider has a frame rate that exceeds 140fps.
4K is where the 3080 lives; this is its bread and butter. It can even hit 60+ frame rates on notoriously GPU-intensive games like Assassin’s Creed: Valhalla. In a wide range of games, the 3070 maintains 60fps 4K averages, but we had to lower settings significantly to maintain those averages.
Tracing by rays
It is likely that you already know that the 30-Series is the pinnacle of hardware ray tracing. After the 20-Series, these 2nd generation RT cores offer twice as much throughput as the 20-Series. In other words, the 30-Series cards don’t take nearly as much of a beating from the RT workload.
In ultra RT mode, lighting, shadows, and reflections appear so realistic, you’ll feel like you’re in the world set forth by your game, but expect to experience some frame drops.
There are 68 RT cores in the 3080, which allows for exquisite lighting in games that support it, but the burden on frames per second can be substantial.
Nvidia was really aiming for a native resolution for each of these cards, so their VRAM specs are entirely different.
Compared to the has 10GB of enhanced GDDR6X memory. With 760.3GBps, the 3080 has nearly twice the bandwidth of the 3070 thanks to its 6X operating frequency. In contrast to the 3070’s 256-bit bus memory, the 3080 has a 320-bit bus memory interface.
Due to the 16-byte burst length of GDDR6, twice that of 6X, the 3070 has the higher 1750MHz memory clock frequency, but 6X offers twice as much efficiency per channel, giving a 19GBps effective speed, 5GBps faster than the 3070. In addition to the extra 2GB storage capacity, this gives the 3080 its high-resolution edge.
If you have a 60Hz or even 140Hz 1440p monitor, the 3070 is the most logical choice and will offer you the highest price per performance ratio.
However, even though we have gone over the RTX 3070 vs 3080 that isn’t the whole story. The bad news is that you’re going to find the GPUs are expensive. Or you may have trouble finding either of these GPUs. It may be worthwhile to wait for the lower end of the RTX 4000 series or the RDNA 3.
Comparison of NVIDIA GeForce RTX 3070 vs 3080
FPS Review: Quick Comparison
High end vs. medium-high. The GPU market has been particularly volatile lately. When these two Ampere architecture GPUs hit the scene in September 2020, there wasn’t much time to compare them before they flew off the shelves. But now there’s time. Is the RTX 3080 faster than the 3070 in terms of gaming performance? There’s a $200 price gap between the MSRPs of the 3070 and 3080, $499 and $699, respectively. Is there any value there?
Here are two NVIDIA contenders to see how much they differ from each other. If you want to learn more about Ampere architecture or take a deeper dive into each GPU individually, read our full RTX 3070 and RTX 3080 reviews. We have guides for overclocking the RTX 3070 and RTX 3080 if you’re interested.
The specifications of the NVIDIA GeForce RTX 3070 and NVIDIA GeForce RTX 3080 (10GB)
Almost every metric of the RTX 3080 outperforms the 3070. Starting with CUDA cores, we see nearly a 50% increase. As a result, the TDP is 100 Watts higher and the boost clock is slightly lower, which is most likely intended to keep it cool. A similar increase in power is seen by the ray tracing cores as well as tensor cores.
Additionally, the 3080 comes with 10GB of memory as opposed to 8GB on the 3070. We also see an evolution to GDDR6X memory, which results in a bit faster performance. As a result, their differences are quite considerable if they are compared to the 3070’s 14GHz memory clock with a 19GHz clock.
Performance in gaming
Here are the results of our previous reviews of these two cards under gaming load:
It’s no surprise that the 3080 wins, but the 3070 holds its own even in some 4K games. However, if you’re going to enable ray tracing, go with the 3080 for 60+ FPS in 4K.
It seems best to choose between 4K and ray tracing with the 3070. If you enable ray tracing, you will be able to go down to 1440p with DLSS. Despite the slightly lower boost clock, the 3080’s extra CUDAs, memory, and RT cores really shine.
Both cards are extremely playable at the highest resolutions with ray tracing, but if you want to be closer to 60 frames per second, the 3080 is a better choice.
It’s time to compare the Nvidia RTX 3070 Ti and 3080.
Both cards are in the mid-range section, so there are a lot of features for you to evaluate which one is best for your needs. Both GPUs are capable of delivering great performance. There is $100 separating their MSRP and the Ti was released about a year later so let’s look at some of their specifications and see how they compare in terms of price vs. performance, to help you decide which one to go with.
We’ve got some good news and some…good news! The good news is that if you’re choosing between these cards, you’re going to have a hard time regardless of which card you choose. Moreover, Nvidia has such a clear purpose for these two cards that you really only have to take into consideration your current configuration to decide which one is best.