RTX 3080 Ti The newest and most powerful card of the series. Here is RTX 3080 Ti Review. RTX 3080 Ti Although it has high performance, it is an expensive graphics card.
The Nvidia GeForce RTX 3080 Ti is the successor to Nvidia’s controversial Turing series and brings improved performance without increasing the price.
When the RTX 2080 * and RTX 2080 Ti * were released, Nvidia caused a lot of excitement by charging a higher price for the two cards – from € 699 to € 1,199 in the case of the GTX 1080 Ti RTX 2080 Ti. That led many to believe that this would happen again this time.
Instead, with the Nvidia GeForce RTX 3080 Ti, we get an improvement of 50 to 80% over the RTX 2080 for the same price and even 20 to 30% over the RTX 2080 Ti, which cost almost twice as much when it was launched.
While we would have been happy if the prices had fallen to the level of the 1000 Pascal series, we can now confirm that this is actually the biggest generational leap in PC graphics that we have seen in years – maybe even at all.
Price and availability
The Nvidia GeForce RTX 3080 Ti will be available from September 17th and starts at € 699 for the Founders Edition. However, as with any major GPU release, there will also be plenty of third-party graphics cards such as MSI, Asus, and Zotac.
It should be noted, however, that some of these third-party cards could be more expensive than the Founders Edition, depending on exotic cooling systems or factory overclocks. But every RTX 3080 Ti should have more or less the same performance as Nvidia’s own card.
Features and chipset
The Nvidia GeForce RTX 3080 Ti is based on the new Nvidia Ampere * graphics architecture, which brings major improvements in both performance and energy efficiency. The fact that Nvidia has improved performance so much over the RTX 2080 and at the same time improved energy efficiency means that overall performance is far superior to any Nvidia Turing graphics card.
Some obvious improvements affect the RT and tensor cores – second and third generation respectively – but the biggest improvement is in the rasterization engine.
Through clever optimization, Nvidia has managed to double the number of CUDA cores in each streaming multiprocessor (SM) by allowing both data paths on each SM floating point 32 (FP32) to process operations – a huge improvement over Turing, where a Data path was only responsible for integer operations. This effectively doubles the raw FP32 throughput per core, even if that doesn’t translate one-to-one to double the frame rate in PC games – at least not for many.
This means that the Nvidia GeForce RTX 3080 Ti has only 46% more SMs than the RTX 2080 with 68, but the number of CUDA cores more than doubled from 2,944 to 8,704. Theoretically, this results in three times the FP32 throughput, from 1ß TFLOPs to 29.7 – an absolutely huge leap from one generation to the next.
When you combine the increased number of CUDA cores with a larger cache, larger texture units, and greater memory bandwidth – thanks to an upgrade to faster GDDR6X memory on a 320-bit bus – gaming performance experiences one of the biggest leaps in years, even if it does fall behind the “double performance” that many had hoped for. But more on that later.
Nvidia RT cores are back – hence the RTX name – and there are huge improvements here too. Nvidia Ampere graphics cards, including the RTX 3080 Ti, have the second generation RT cores that function similarly to the first generation but are twice as efficient.
With ray tracing, the SM throws a light beam in the scene that is being rendered, and then the RT core takes over, makes all the calculations about the reflections of the light beam, and sends this information back to the SM. This means that the SM can calmly render the rest of the scene. But we have still not reached the point where ray tracing does not lead to a loss of performance. Maybe one day.
Tensor cores are now twice as powerful, which is why Nvidia only installed four instead of eight in each SM, as was the case with a Turing SM. Because there are generally more SMs, the DLSS performance also increases immensely.
This generation of graphics cards is not just about gaming, because Nvidia also has a few features up its sleeve that make life easier for everyone who owns an RTX card.
For example, we’re already big fans of RTX Voice and Nvidia has finally brought the feature out of beta into a comprehensive streaming app. While RTX Voice filters out background noise from your microphone, the broadcaster can filter out backgrounds or even blur them.
The video section is still in beta and we actually ran into a few glitches, but it’s already way better than any other solution that lets you replace the green screen background.
One of the features we’ve been waiting for the most, however, is Nvidia RTX I / O, an API that works with Microsoft’s DirectStorage API and transfers data directly from your SSD to the graphics card. In the case of next-gen games, this should not only drastically reduce loading times but also “copy” the revolutionary I / O performance of the PS5 and Xbox Series X.
Unfortunately, this technology has to be built into their games by developers, so we couldn’t test what difference it actually makes. However, we will keep an eye on it and test it once it is widely implemented. And since consoles will use similar technology, we assume it will arrive faster than ray tracing.
As for the actual Founders Edition of the graphics card, Nvidia relies on a completely new fan design that is more practical than any reference design it has ever designed. The company uses a shorter, multilayer circuit board to allow the heat sink to occupy the entire bottom end. This enabled Nvidia to install a fan on the back that sucks cold air into the heat sink and blows it out again at the other end of the case.
The system we used to test the Nvidia GeForce RTX 3080 Ti Founders Edition:
CPU: AMD Ryzen 9 3950X (16 cores, up to 4.7 GHz)
CPU cooler: Cooler Master Masterliquid 360P Silver Edition
RAM: 64 GB Corsair Dominator Platinum @ 3,600 MHz
Mainboard: X570 Aorus Master
SSD: ADATA XPG SX8200 Pro @ 1 TB
Power supply: Phanteks RevoltX 1200
Case: Wetbench practice
From Nvidia’s (exaggerated) marketing alone, we assumed that the RTX 3080 Ti would become a fast graphics card, but “fast” is an understatement. From the moment we took it out of the box, it stuck in our own PC and packed everything from Final Fantasy XIV to Control and just came out of the personal tower to be plugged into the test PC.
Before 3080, we used the Nvidia GeForce RTX 2080 Ti in the PC and the difference was already noticeable before we could significantly measure the performance. In one of the games we play the most, Final Fantasy XIV, the RTX 2080 Ti slumped below 60 FPS at 4K, especially since the last expansion. That doesn’t happen with the RTX 3080 Ti. This means that the game runs somewhere between 75 and 100 FPS on the highest graphics settings – a huge jump for only half the price.
It went on and on, no matter what game. Metro Exodus on highest settings with ray tracing and DLSS? Smooth in 4K with 60 FPS locked. Control with countless ray tracing effects? Soft as butter. Even final Fantasy XV with all the strange optional graphic effects runs at 4K with around 60 FPS. 4K60 gaming has arrived in the mainstream – even if we stretch the “mainstream” a little.
While the heat values in our tests are not particularly exciting because they took place with the case open, the temperatures in our personal, closed tower settled at 60 ° C – well below the temperatures in the 80s range of the RTX 2080 Ti Founders Edition.
When looking at the actual benchmark results, it becomes clear that the Nvidia GeForce RTX is in a different league and is itself well above the RTX 2080 Ti. In 3DMark Time Spy Extreme, the RTX 3080 Ti is a whopping 63% faster than the RTX 2080 and 26% faster than the 2080 Ti – a huge leap when you consider that the RTX 2080 was only 40% faster than the GTX 1080 than we were they tested in 2018.
But Time Spy Extreme is not even the best scenario for the power of the RTX 3080 Ti. In Red Dead Redemption 2, where we had all graphics settings except MSAA – multi-sample anti-aliasing is extremely demanding and not worth it – we were able to do one See an insane 87% increase compared to the previous generation.
This doesn’t quite come close to twice the performance promised at the RTX 3080 Ti launch, but it is closer than we would have thought. The RTX 3080 Ti is overall between 50 and 80% faster than the RTX 2080 and only falls a little below in Fire Strike Ultra, where it only achieves a lead of 29% – which is still a decent lead.
The big difference in performance is only noticeable in 4K, however, if the graphics cards are not affected by other bottlenecks. For many titles, even the setup of an AMD Ryzen 9 3950X and 64 GB RAM at 3,900 MHz limited the RTX 3080 Ti. This is also the reason why the RTX 2080 Ti and RTX 3080 Ti are practically on par with Metro Exodus in 1080p and an advantage of 19% arises at 4K.
Therefore, we can only recommend the Nvidia GeForce if you also want to play in 4K, or maybe in 3,400 x 1440 ultrawide. At lower resolutions, the performance difference will simply not be noticeable, so that you would be better served with the GeForce RTX 3070
In the upper area, thanks to the Nvidia GeForce RTX 3080 Ti, the lead over AMD continues to grow and in many tests, it is already twice as large compared to AMD’s most powerful consumer graphics card, the Radeon RX 5700 XT. AMD Big Navi has to bring a lot to the table to dust off the 4K crown that the RTX 3080 Ti has just won.
The Nvidia GeForce RTX 3080 Ti is an absolute powerhouse of a graphics card and brings one of the biggest leaps in performance in GPU history. Anyone interested in 4K gaming should look out for this graphics card – even if the benefits are less noticeable at lower resolutions.
- Excellent 4K gaming performance
- Low temperatures
- Lots of useful non-gaming features
- Price not increased compared to RTX 2080
- Still pretty expensive
- The dongle is a bit annoying
- Founders Edition without USB-C