GeForce RTX 2080 Ti 11 GB: Overview| Test | Specs| Is it Good for Gaming (Reaching 40%) – The most powerful graphics card of 2018 was announced by NVIDIA on August 20. The novelty was named GeForce RTX 2080 Ti, and despite belonging to the GeForce series, it is positioned by the manufacturer as a graphics adapter of the level of video cards of the TITAN line of previous generations. This is also reminiscent of the cost of the model – $ 1199. That’s how much NVIDIA asked for the latest TITAN Xp.
NVIDIA RTX Technology
For those who are not familiar with recent events, let’s say that the top video cards of the new 20th GeForce series have received the name RTX, instead of the former GTX. This change is due to the introduction of the ray tracing technology of the same name into the Pascal graphics architecture. NVIDIA RTX. You can learn more about the new method of graphic processing from this article. Speaking very briefly about the new technology, it allows you to bring the visual component of many computer games to a whole new level.
MSI RTX 2080 Ti GAMING X TRIO 11GB Unboxing & Benchmark
This applies to the reflection of models, the refraction of light, dimming, the appearance of shadows and many other points. The effect of this technology can be clearly seen below on the demo video of the game Metro Exodus. Currently, game creators are already adopting the new technology from NVIDIA, and the first games with the RTX function will appear very soon.
GeForce RTX 2080 Ti 11 GB: Specifications
The basis of the GeForce RTX 2080 Ti video adapter is a graphics processor TU102. This GPU received tensor cores for deep learning operations and the operation of the DLSS anti-aliasing mode, as well as RT cores for updated ray tracing.
Let’s compare the characteristics of the novelty with its predecessor GeForce GTX 1080 Ti by Founders Edition.
| GeForce RTX 2080 Ti | GeForce GTX 1080 Ti | |
| GPU model (architecture) | TU102 (Turing) | GP102 (Pascal) |
| Technical process | TSMC 12 нм | TSMC 16 нм |
| Number of transistors | 18.6 billion | 12 billion |
| CUDA stream processors | 4352 | 3584 |
| Number of Tensor cores | 544 | – |
| Number of RT cores | 68 | – |
| texture blocks | 272 | 224 |
| Render Units (ROPs) | 88 | 88 |
| core frequency base | 1350 MHz | 1480 MHz |
| Boost Core Clock | 1635 MHz | 1582 MHz |
| Memory bus | 352 pages | 352 pages |
| Memory frequency | 1750 MHz | 1375 MHz |
| Memory | 11 GB | 11 GB |
| Memory type | GDDR6 | GDDR5X |
| Bandwidth | 616 GB/s | 484 GB/s |
| power usage | 260 W | 250 W |
| Min. system requirements | 650 W | 600 W |
| Maximum GPU temperature | 89 °С | 91 °С |
| Interface | PCI-E 3.0 | PCI-E 3.0 |
| Add. nutrition | 2x 8-pin | 1x 8-pin + 1x 6-pin |
| Communication | 2x NVLink | SLI |
In addition to the tensor and RT cores that appeared in Turing graphics cards, the new product has 4352 CUDA stream processors, which are 21% more than the GeForce GTX 1080 Ti. And the transition to the new GDDR6 video memory standard increased the bandwidth to 616 GB / s. As for the amount of memory, it remained the same as that of its predecessor – 11 GB.
Of the innovations of the Turing architecture, we highlight the anti-aliasing mode DLSS (the name is an abbreviation of the term Deep Learning Super Sampling). The specified algorithm is a logical development of the TAA anti-aliasing method, due to the fact that it does not always work perfectly and sometimes causes artifacts. Unlike TAA, Deep Learning Super Sampling mode allows you to display games in 4K with the maximum level of detail, at the same time with higher FPS and without annoying jerks. The only caveat is that the DLSS method is an innovation and requires its implementation into games directly by their developers. An example of the operation of this algorithm can be seen on the example of the game Final Fantasy XV.
In this section, we can also note the updated power subsystem for GeForce RTX series video cards, in the development of which NVIDIA, by its own admission, has invested a lot of time and money. The new subsystem provides noticeably more efficient operation, higher clock speeds than before, and at the same time has higher overclocking potential.
Appearance
The Founders Edition video card received a casing made of anodized aluminum. A black panel was placed in front between the fans. The back of this model is adorned with an LED illuminated logo. On the side, a removable cover covers the NVLink connector. In terms of weight, new video cards are heavier than previous generation cards.
Instead of the usual “turbine”, the cooling system of the novelty received a pair of 85 mm axial fans. Two fans are new for NVIDIA’s FE models, but quite successful at the same time. In idle mode, the fan speed is at the level of 1500 rpm, and under load it reaches 2000 rpm. The new CO has a noise level a few decibels lower than previous Pascal architecture video adapters. Of the minuses, one can note the continuous operation of the fans even when the video card is idle.
Two 8-pin sockets supply additional power to the video card. The TDP level of the GeForce RTX 2080 Ti model is 260 W, while the user receives about 60 W for overclocking.
The classic version has one HDMI 2.0b video output and three DisplayPorts 1.4. We also note that NVIDIA became the first company to install an interface VirtualLink. This connector looks like a Type-C port, and thanks to it it became possible to connect virtual reality helmets to a computer with a single cable. Despite the fact that the port looks like USB Type-C, a completely different protocol is used for data transfer.
Testing the GeForce RTX 2080 Ti in games
For testing, the stand configuration was used:
- processor Intel Core i7-6850K,
- motherboard MSI X99S MPower,
- memory G.Skill F4-3200C14Q-32GTZ (32 GB),
- SSD Plextor PX-256M6Pro (256 MB),
- Corsair AX1500i power supply (1500 W),
- operating system Windows 10 Pro x64,
- GeForce 416.34 WHQL driver.
Games that support NVIDIA RTX technology are just getting ready for release at the time of testing. Accordingly, the tests were carried out without using this technology with a resolution of 2560×1440 and 3840×2160. The screenshots below show the graphics settings and the average number of received fps in each game.
Conclusion
Comparing single-class models of the latest generations, we see a performance advantage of the GeForce RTX 2080 Ti over the GeForce GTX 1080 Ti reaching 40%. This is at its maximum. In the minimum value in some games, the new video card gives an increase of no more than 15%. If we compare the new product with the previous flagship GeForce GTX 980 Ti, then we get a real twofold superiority in performance.
The new flagship from NVIDIA allows you to run absolutely all the games that existed at the time of its release with a stable 60 FPS at 4K resolution. True, sometimes with a little adjustment of the settings. As for the NVIDIA RTX function, we can talk about this a little later, when at least a few game projects with support for this option are released, and when it becomes possible to test these games “live”.
Most likely in 2018, the model considered in this review will remain the top. But next year, it is quite possible that NVIDIA will announce something new, like “TITAN RTX”, on a full-fledged, not “stripped down” version of the TU102 core.
It is clear that the top Turing video cards are designed for a small segment of the market due to their high cost, and a significant part of gamers are waiting for the release of more affordable NVIDIA solutions. In the near future, we will consider the junior model of the new GeForce family – the RTX 2070 video card, and we will wait for the release of the mid-range 2060 and 2050 graphics adapters. Unfortunately, their announcement will take place only in 2019.
