Nvidia's GeForce RTX 2080 and RTX 2080 performance digests, features and prices

[ad_1]
<div _ngcontent-c15 = "" innerhtml = "

Yesterday, Nvidia unveiled the performance of its new graphics cards GeForce RTX 2080 and GeForce RTX 2080 Ti, based on the microarchitecture GPU Turing society. If you've followed the launch over the past month, you're probably aware that Turing not only brings additional GPU resources for traditional rasterization rendering, but also advanced technologies with new core engines for ray tracing. real time. and even Tensor cores for in-depth learning and AI-assisted rendering techniques.

You are probably also aware that the GeForce RTX 2080 and 2080 Ti cards that were launched are relatively expensive compared to Nvidia's previous generation high-end graphics cards, such as the GeForce GTX 1080 and 1080 Ti. Specifically, a GeForce RTX 2080 Founder Edition card has a MSRP of $ 799, while the GeForce RTX 2080 Ti FE is priced at $ 1,199. In its debut, Nvidia's GeForce GTX 1080 Ti had an MSRP of $ 699, so price deltas are definitely significant. As a result, high expectations have been raised by the performance evaluations that have been launched.

HotHardware compared these two Nvidia Turing-based cards in a wide range of GPGPU gaming engines and GPGPU workloads, as well as a myriad of technology sites on the Web. HotHardware noted that the performance profiles were functioning as follows: "The GeForce RTX 2080 Ti is approximately 20 to 45% faster than a Titan Xp and makes the game 4K smooth with very high image quality settings. high with a single GPU. The performances were particularly strong with the new optimized titles such as FarCry 5 and the benchmarks related to virtual reality, but the card dominated all the game tests. "Regarding the GeForce RTX 2080, the L & # 39; edition Founder's Edition of Nvidia or the custom EVGA and MSI cards that we have also presented are a little more difficult to summarize. The GeForce RTX 2080 far exceeds the GeForce GTX 1080 by a wide margin. His performances were also good in new titles and tests related to virtual reality, and have generally reached a level comparable to that of a GeForce GTX 1080 Ti, even if they have exceeded the Titan Xp twice.

So it's long and short on the performance, although you can feel safe from the numbers, of course. Whether or not these performance figures make sense, compared to the price premium, the RTX 2080 and RTX 2080 Ti order is where the rubber meets the demand of many players and enthusiasts of the Web. The analysis and gossip are ongoing, with some citing that a gain of 20 to 45% for a $ 1,200 graphics card is exorbitant. Meanwhile, lower gains or parity performance in the case of the $ 800 RTX 2080, against a 1080 Ti that can be obtained for only $ 150 less, are not enough. However, I would suggest that Nvidia put its bets on the next-generation Turing graphics rendering engines that I mentioned earlier, and that gamers and graphics professionals can achieve more performance and features back on investment in these two Nvidia Turing GPUs.

It's not easy to be a pioneer, and admit that any new generation of product in a family of graphics cards carries an inherent risk. However, what Nvidia did with Turing is nothing less than true gaming graphics innovations when you consider features such as DLSS or Deep Learning Super Sampling and its RT cores for time ray tracing. real. These blocks occupy much of the silicon real estate in Turing, and the silicon surface, pure and simple, is directly related to costs. There are not two ways to get around this. As a reference, the Pascal GPU on a GeForce GTX 1080 Ti measures 471 square millimeters, while the TU102 on the RTX 2080 Ti measures 754 square millimeters. This represents a massive 60% increase in the field and a huge commitment from Nvidia for the new core resources of Tensor and RT. To be honest, all this extra area is not devoted to these new processing engines, but you can bet that a big part of this area is. Now, investing in software tricks for a next-generation graphics card is one thing, but do you think Nvidia would invest large blocks of silicon real estate (and a cost as a direct result) in the resources and functionality of the GPU? Would not trusting provide significant tangible benefits?

If you've seen DLSS in action in the last few weeks and you've seen the performance benefits and image quality it brings, you know that Turing could achieve much greater performance gains. DNN (or Deep Neural Network) models are designed for game engines on Nvidia servers, and then forwarded to the Turing GPU of the client machine to update a game library in Nvidia's GeForce Experience software suite. Instead of having to sample and render an image again and again to reduce folding, the algorithm learned by a machine simply uses the model to make the images as clear as possible. For me, this is an excellent use of machine learning inference to improve performance and visuals, and Nvidia uses Turing Tensor cores to achieve this.

In addition, if you've seen real-time raytracing in action, you'll know the increased level of realism for gaming and virtual reality applications that properly rendered lighting, shadows, and reflections can bring. In mixed rendering, with traditional rasterization and combined ray tracing, this can help make a scene more cinematic and realistic. Real-time ray tracing has been designated as the "holy grail" of game graphics and other applications, and you also need dedicated hardware like RT kernels. It will work on existing hardware without specific RT resources, of course, if you like to watch slideshows.

So, before you decide too quickly on the cost of Nvidia's new GeForce RTX 2080 and RTX 2080 Ti, remember the intrinsic cost of innovation. I would not want to justify the cost of these new graphics cards in relation to your own scale of value. It's up to you to decide. In addition, game developers will need to embrace these new features and techniques, some of which are easier to implement than others. However, I have been covering computer graphics for almost two decades and I would say that Nvidia's investment in silicon for next-generation graphics technologies like DLSS and ray tracing will be faster than later for PC gamers. and the professionals of the workstations.

">

Yesterday, Nvidia unveiled the performance of its new graphics cards GeForce RTX 2080 and GeForce RTX 2080 Ti, based on the microarchitecture GPU Turing society. If you've followed the launch over the past month, you're probably aware that Turing not only brings additional GPU resources for traditional rasterization rendering, but also advanced technologies with new core engines for ray tracing. real time. and even Tensor cores for in-depth learning and AI-assisted rendering techniques.

You are probably also aware that the GeForce RTX 2080 and 2080 Ti cards that were launched are relatively expensive compared to Nvidia's previous generation high-end graphics cards, such as the GeForce GTX 1080 and 1080 Ti. Specifically, a GeForce RTX 2080 Founder Edition card has a MSRP of $ 799, while the GeForce RTX 2080 Ti FE is priced at $ 1,199. In its debut, Nvidia's GeForce GTX 1080 Ti had an MSRP of $ 699, so price deltas are definitely significant. As a result, high expectations have been raised by the performance evaluations that have been launched.

HotHardware compared these two Nvidia Turing-based cards in a wide range of GPGPU gaming engines and GPGPU workloads, as well as a myriad of technology sites on the Web. HotHardware noted that the performance profiles were functioning as follows: "The GeForce RTX 2080 Ti is approximately 20 to 45% faster than a Titan Xp and makes the game 4K smooth with very high image quality settings. high with a single GPU. The performances were particularly strong with the new optimized titles such as FarCry 5 and the benchmarks related to virtual reality, but the card dominated all the game tests. "Regarding the GeForce RTX 2080, the L & # 39; edition Founder's Edition of Nvidia or the custom EVGA and MSI cards that we have also presented are a little more difficult to summarize. The GeForce RTX 2080 far exceeds the GeForce GTX 1080 by a wide margin. His performances were also good in new titles and tests related to virtual reality, and have generally reached a level comparable to that of a GeForce GTX 1080 Ti, even if they have exceeded the Titan Xp twice.

So it's long and short on the performance, although you can feel safe from the numbers, of course. Whether or not these performance figures make sense, compared to the price premium, the RTX 2080 and RTX 2080 Ti order is where the rubber meets the demand of many players and enthusiasts of the Web. The analysis and gossip are ongoing, with some citing that a gain of 20 to 45% for a $ 1,200 graphics card is exorbitant. Meanwhile, lower gains or parity performance in the case of the $ 800 RTX 2080, against a 1080 Ti that can be obtained for only $ 150 less, are not enough. However, I would suggest that Nvidia put its bets on the next-generation Turing graphics rendering engines that I mentioned earlier, and that gamers and graphics professionals can achieve more performance and features back on investment in these two Nvidia Turing GPUs.

It's not easy to be a pioneer, and admit that any new generation of product in a family of graphics cards carries an inherent risk. However, what Nvidia did with Turing is nothing less than true gaming graphics innovations when you consider features such as DLSS or Deep Learning Super Sampling and its RT cores for time ray tracing. real. These blocks occupy much of the silicon real estate in Turing, and the silicon surface, pure and simple, is directly related to costs. There are not two ways to get around this. As a reference, the Pascal GPU on a GeForce GTX 1080 Ti measures 471 square millimeters, while the TU102 on the RTX 2080 Ti measures 754 square millimeters. This represents a massive 60% increase in the field and a huge commitment from Nvidia for the new core resources of Tensor and RT. To be honest, all this extra area is not devoted to these new processing engines, but you can bet that a big part of this area is. Now, investing in software tricks for a next-generation graphics card is one thing, but do you think Nvidia would invest large blocks of silicon real estate (and a cost as a direct result) in the resources and functionality of the GPU? Would not trusting provide significant tangible benefits?

If you've seen DLSS in action in the last few weeks and you've seen the performance benefits and image quality it brings, you know that Turing could achieve much greater performance gains. DNN (or Deep Neural Network) models are designed for game engines on Nvidia servers, and then forwarded to the Turing GPU of the client machine to update a game library in Nvidia's GeForce Experience software suite. Instead of having to sample and render an image again and again to reduce folding, the algorithm learned by a machine simply uses the model to make the images as clear as possible. For me, this is an excellent use of machine learning inference to improve performance and visuals, and Nvidia uses Turing Tensor cores to achieve this.

In addition, if you've seen real-time raytracing in action, you'll know the increased level of realism for gaming and virtual reality applications that properly rendered lighting, shadows, and reflections can bring. In mixed rendering, with traditional rasterization and combined ray tracing, this can help make a scene more cinematic and realistic. Real-time ray tracing has been designated as the "holy grail" of game graphics and other applications, and you also need dedicated hardware like RT kernels. It will work on existing hardware without specific RT resources, of course, if you like to watch slideshows.

So, before you decide too quickly on the cost of Nvidia's new GeForce RTX 2080 and RTX 2080 Ti, remember the intrinsic cost of innovation. I would not want to justify the cost of these new graphics cards in relation to your own scale of value. It's up to you to decide. In addition, game developers will need to embrace these new features and techniques, some of which are easier to implement than others. However, I have been covering computer graphics for almost two decades and I would say that Nvidia's investment in silicon for next-generation graphics technologies like DLSS and ray tracing will be faster than later for PC gamers. and the professionals of the workstations.