Nvidia announced their next generation Graphics during their special GeForce gaming event, where they talked about some of the new features exclusive to these cards only. We have already talked about the feature driven by AI called DLSS where the GPU will decide what resolution will better suit the game and enhances the resolution of the image through their trained neural networks. Secondly, there is real-time raytracing introduced by the new RT cores, the sole purpose of them is to render the light following its physics and not through traditional rasterization. We have already talked about these; you can follow the link to read about them.
The topic at hand today is the early pros and cons of the upcoming Graphics card based on the information we have as the Graphics cards are not on the market yet. So, try to bear with me and digest some of the information with a grain of salt. These new graphics cards represent a world where Nvidia has a monopoly over the high-end market with no competition, and the result is offensively high prices. The other advantage Nvidia has due to market dominance is the authority to mark its flag of new “not so welcomed” features. Keeping these in mind lets get right into the details.
Pro: Factory overclocked Founder’s edition cards
Since the start of the Founder’s edition cards, the trend was to make an aesthetically pleasing card with the stock features and blower style cooler so that they can fit into mini builds. But the main purpose of the Founder’s edition card is that they are the reference hardware for the partner firms to make their Video cards. The reason why the FE Graphics cards were not that popular and only the Nvidia “Fanboys” would have these Graphics cards on board was the extra premium that they charge and no added benefits. The partner firms like Zotac, ASUS, Gigabyte, etc. have much better factory overclocked with better cooling solutions at a lower price.
This time around Nvidia is giving perks with the FE Graphics cards as they are now shipped with a factory overclocked GPU inside. The cooling solution now has a dual fan design which will definitely have better outcomes than the blower style fan on the previous Founder’s edition cards. The dual fan design will also help in providing enough cooling to overclock the already overclocked GPU making overclocking a better suite for the FE cards. Overclocked FE cards also mean that the reference Graphics cards will come from the partner firms and the initial reviews will have better performance charts.
Con: Super High Price
One of the critiques that the Founder’s edition Graphics cards usually face since their inauguration. The reason behind the premium is the shiny materials used to build the exterior of the Graphics cards and the satisfaction to be one of the first users of the new technology. This time though the package is better than the previous generations due to Factory overclocked GPUs and better cooling solution but the fact that many upcoming Graphics cards will have same features for less.
Even if we do not consider the FE Graphics cards the prices introduced by Nvidia are almost double than what the respective predecessor is retailing at the moment. There is always a premium for the generation, but Nvidia is going all out this time around. The GTX 1080Ti was launched with a price of 649 dollars while its Turing successor called RTX 2080Ti has a launch price of 999 dollars. The price is 64% higher while the performance difference between both Graphics cards is roughly 35 percent to 45 percent.
Which makes upgrading from 1080Ti to 2080Ti is a rather bad deal, however, if you are upgrading from a GTX 1080Ti to RTX 2080, then the price difference is not huge, and same goes for the performance. You will only be getting the benefits of Raytracing and DLSS with this deal.
Pro: Real-time Raytracing
The focus point of the presentation during the GeForce event was the Raytracing abilities of the new Turing architecture. We can see that Nvidia is wholeheartedly banking on the Raytracing this time since the performance difference is not as huge as the Maxwell and Pascal performance difference was. The reason behind the little (It’s not little, In comparison, it is not that huge) performance difference can be the fabrication jump which was too little this time.
If we see the Maxwell 2 Graphics cards, they were made with 28nm fabrication process while the initial Pascal graphics cards were made with 16nm and then 14nm for the low-end devices. But the Turing architecture is made with a 12nm TSMC’s FinFET fabrication process. The generation gap is actually negligible the 12nm process is actually an improved version of the 16nm process, it does not have the full-fledged advantages of a denser process rather improvements of the qualities build by the 16nm process. That is why the raytracing is important in the scenario and the fact that the RTX 2080Ti can trace 10 BILLION rays of light in one second with all the physical properties that light possess. The GTX 1080Ti can only trace a billion with rasterized rendering.
Real-time Raytracing is called the Holy Grail of graphics which makes it a big deal but the implementation in games is questionable as the developers would have to employ the Raytracing in their games.
Con: Performance during Raytracing
When you are paying $1200 for a Graphics card, it is your right to play each game at highest possible settings at 4K resolution at the very least 60 FPS. That is not going to be the case If you through in real time Raytracing in the mixture then the 4K performance along with handling billions of light rays at the same time would be impossible.
At the recent event where Nvidia launched these Graphics cards two of the most anticipated games Shadow of Tomb Raider and the Battlefield V were showcased, with RTX enabled. Even though the games were looking absolutely stunning but the resolution they are running at is not 4K, but the GPU these games are using a full GeForce RTX 2080Ti. The actual resolution of these games is 1080p with framerates closing to 60 for the Battlefield V and around 40s for the Shadow of the Tomb Raider. While the Graphics card was handling all rays efficiently in Battlefield V but even at low shadows, the RTX 2080Ti could not hit 60FPS at 1080p. It shows that rendering real light particles is a taxing task for the GPU even though there are separate cores for raytracing.
It can be argued that these games are the Alpha builds and due to fewer optimizations, these games are not utilizing the hardware fully. We can see the performance of the raytraced GPU only at the time when we have the release versions of the game. It cannot be said with certainty what kind of performance we will be getting from the RTX graphics cards on a raytraced game.
Pro: Artificial intelligence in gaming
A few years ago, even a month ago artificial intelligence only had used in a corporate world where the Graphics giant Nvidia was making loads of money because of the monopoly it has on the market. Now with the Turing architecture, Nvidia is introducing AI in gaming along with raytracing.
Unlike Raytracing inclusion of AI in graphics is a totally new concept and Nvidia is confident about it not as confident as in the case of Raytracing. Nvidia is adding Raytracing into the gaming through its image enhancing technique called DLSS (Deep learning supersampling). The technique is (simple at least in theory) that Nvidia will train a neural network through its high powered supercomputer called Saturn V, the purpose of the neural network is to convert a lower resolution picture to a high definition image. The trained neural networks will be added to the systems of the consumers through recursive software updates.
Con: Adoption and mainstream graphics
The real problem with the new features is their adoption rate. Currently, only a handful of games have adopted DLSS and Real-time raytracing since it requires a full-time dev to include these features in the games. The gaming market nowadays is mostly focused on the consoles since more than 60 to 80 percent of the games are played through the console platform, and PC also has pirated games which are also discouraging for the developers.
The raytracing abilities showcased by the top graphics card are just enough for the games even an RTX 2070 would output a beautifully rendered raytraced image, but what about the low-end SKUs. Most of the PC community uses low-end or mid-end hardware if the RTX 2070 is the absolute minimum then will there be any other graphics card with an RTX prefix is the question since real-time ray tracing will be out of the equation. Lastly, the game development would require the assistance of Nvidia which is a fully corporate topic and some of the low-end developers would not get a chance to add raytracing in their games. The impacts of DLSS will be huge since the game will actually be running at a lower resolution but the image output will be a lot higher, and thus the better will be far better.
These are some of the pros and cons that you can expect if you are closely following the RTX Graphics cards. There may be some changes in them when the actual cards hit the market, and we see actual benchmarks and gameplay. So stay tuned for more updates on the topic.