Nvidia rtx 【all information】

We already have with us the new NVIDIA RTX graphics cards. From the flagship model: NVIDIA RTX 2080 Ti, to the model for the most gamers in 4K: NVIDIA RTX 2080 and the one that is most affordable for all budgets, the NVIDIA RTX 2070. In this article we will explain what are its novelties and new technologies.

Prepared? Let's start!

Index of contents

We summarize the best hardware guides that you are surely interested in reading:

• Best processors on the market Best motherboards on the market Best RAM memory on the market Best graphics cards on the market Best SSDs on the market Better chassis or PC cases Better power supplies Better heatsinks and liquid coolers

Ray Tracing more present than ever

Ray Tracing is one of the most talked about terms since the arrival of Nvidia GeForce RTX graphics cards, as they are the first in history that are capable of applying this technology in real time to video games. Nvidia's Ray Tracing implementation is called RTX, hence this is the new suffix for the company's graphics cards. But what is Ray Tracing and RTX technology? We have prepared this post to explain the fundamentals of these new technologies and graphics cards.

There may not be many people outside of computer graphics who know what Ray Tracing (also known as ray tracing) is, but there are very few people on the planet who have not seen it. Ray Tracing is the technique on which modern films are based to generate or improve special effects. Think of realistic reflections, refractions, and shadows. This makes the starfighters in sci-fi epics scream, the fast cars look furious, and the fire, smoke, and explosions of war movies look real.

It also produces images that may be indistinguishable from those captured by a camera. Live-action movies mix computer-generated effects and real-world images captured seamlessly, while animated films cover digitally generated scenes in light and shadow as expressive as anything shot by a cameraman. The easiest way to think about Ray Tracing is to look around you. Right now, the objects you are viewing are illuminated by rays of light from the sun. Now turn around and follow the path of those rays backwards from your eye to the objects with which the light interacts. That is ray tracing or Ray Tracing.

We recommend reading our post on How to improve the graphic quality of games through supersampling

Historically, PC hardware has not been fast enough to use these techniques in real time in video games. Filmmakers can take as long as they want to render a single frame, so they do it offline on rendering farms. Video games are only a fraction of a second, as a result of the inability to use Ray Tracing, most real-time graphics are based on another technique, rasterization.

NVIDIA RTX is Nvidia's implementation of Ray Tracing in video games thanks to Turing

As GPUs continue to become more powerful, ray tracing will work for more and more people in the next logical step in this technology. For example, armed with professional ray tracing tools, product designers and architects use Ray Tracing to generate photorealistic models of their products in seconds, allowing them to better collaborate and omit expensive prototypes. Ray Tracing has proven its effectiveness to lighting architects and designers, who are using its capabilities to model how light interacts with their designs.

GPUs offer more and more power, making video games the next frontier for this advanced technology. In August Nvidia announced its new GeForce RTX graphics cards based on the Turing architecture and compatible with Ray Tracing in real time thanks to RTX technology. It is the result of a decade of work on computer graphics algorithms and GPU architectures.

Nvidia's RTX technology consists of a ray tracing engine that runs on GPUs with Turing or Volta architecture. Designed to support ray tracing through a variety of interfaces, Nvidia partnered with Microsoft to enable full RTX support through Microsoft's new DirectX Ray Tracing (DXR) API. To help game developers take advantage of these capabilities, Nvidia also announced that the GameWorks SDK will add a crawl reduction module. The updated GameWorks SDK, coming soon, includes ray traced area shadows and bright reflections with Ray Tracing. DXR fully integrates ray tracing into DirectX, allowing developers to integrate ray tracing with traditional rasterization and calculation techniques.

Nvidia is developing a Ray Tracing extension for Vulkan's multiplatform graphics and computing API. This extension will be available soon and will allow Vulkan developers to access the full power of RTX. Nvidia is also contributing the design of this extension to the Khronos Group as a contribution to potentially bring inter-vendor lightning tracking capability to the Vulkan standard.

All of this will give game developers the ability to incorporate ray tracing techniques into their work to create more realistic reflections, shadows, and refractions. As a result, the games you enjoy at home will reap more of the cinematic qualities of a Hollywood blockbuster.

Turing, the new graphic architecture

For now only three graphics cards based on Nvidia's Turing architecture have been released, these are the GeForce RTX 2080Ti, RTX 2080 and RTX 2070. Turing is the most advanced graphics architecture of Nvidia, it is an evolution of Volta in which all the benefits of this have been maintained, and new units dedicated to Ray Tracing have been added. These dedicated Ray Tracing units are the RT cores, thanks to which Turing can be up to 10 times more efficient than Volta when working with raytracing.

The Turing power is still insufficient to use Ray Tracing very intensively, which is why only a small amount of light rays is applied. This causes the appearance of an image with a lot of noise, something that nobody likes. This is where the Tensor Core comes into the picture, which is also present in Turing and has the function of speeding up the GPU's artificial intelligence operations. Thanks to these Tensor Core, the GeForce RTX apply advanced algorithms to eliminate image noise and offer an unprecedented level of graphic quality, very similar to what would be obtained with a much more intensive use of raytracing.

Turing's benefits go far beyond Ray Tracing, as this architecture is also a breakthrough against Pascal in every detail. Pascal is the architecture that Nvidia has used in the gaming sector before Turing, since Volta has not reached the world of video games.

The Turing architecture introduces profound changes at the level of SM units (streaming multiprocessors), this is the minimum functional unit of the Nvidia architecture, which includes the CUDA Core, The Tensor Core, the load / save units, and a cache of level 0. For now it is not known if the RT cores are also within the SM, although the logical thing is to think that they are.

Within each SM is also the L1 cache, which in the case of Turing is 128 KB, just like Volta. This cache is responsible for saving the data that is most used by CUDA cores, as well as not being consistent, which means that there is no synchronization between the data in the L1 cache of each SM unit. This L1 cache makes a big difference, since before Turing there was a second memory that was coherent and unified. Turing combines the L1 cache and that second memory into a single inconsistent pool. This will give developers greater flexibility of use, allowing for more optimization as long as they are willing to spend more time on development.

This unification of memory in Turing offers a greater bandwidth and a greater speed at the time of moving the data between this memory and the registers of the CUDA cores. This reduction in access time translates into less need for clock cycles to execute operations in the CUDA Core. Nvidia has stated that the performance of each Turing CUDA core is 50% higher than in Pascal, without a doubt the internal modifications of the architecture have paid off.

Another important change of Turing against Pascal we see in the L2 cache, which has doubled from 3 MB to 6 MB for each SM. Caching is expensive to implement, so its duplication makes it very clear that Turing cores are more powerful than Pascal cores and need more of this precious resource. The L2 cache is where the data that does not fit in the L1 cache is stored, a greater amount means being able to store more data, so less access to the VRAM memory of the graphics card will be needed, translating into a lower consumption of quantity of this memory and energy.

This is important because the Nvidia GeForce RTX have not increased the amount of VRAM compared to Pascal, although the jump has been made to the GDDR6 that offers better energy efficiency and greater bandwidth. This greater bandwidth will allow Turing to perform better than Pascal in high resolutions, so we could finally be before the first graphic architecture that allows taking advantage of 4K G-Sync HDR monitors in all their splendor.

The greater bandwidth of the GDDR6 memory and the lower consumption of this thanks to the improved Turing cache, allows the bandwidth of the cards to be adequate for the correct operation of the RTX technology, since there is a lot of information that the card has to move.

Nvidia RTX models

The following table summarizes the features of Turing-based cards that have been announced to date:

Nvidia GeForce 2000 series
	Silicon	CUDA Core	Giga Rays / s	RTX-OPS	GPU frequency	Memory	Interface	Bandwidth	TDP
Nvidia GeForce RTX 2080Ti	TU102	4352	10	78T	1635 MHz	11 GB GDDR6	354 bits	616 GB / s	260W
Nvidia GeForce RTX 2080	TU104	2944	8	60T	1545 MHz	11 GB GDDR6	256 bits	448 GB / s	225W
Nvidia GeForce RTX 2070	TU104	2304	6	Four. Five	1710 MHz	8 GB GDDR6	256 bits	448 GB / s	175W

The landing of the rest of the Nvidia GeForce 2000 series graphics cards will be completed over the coming weeks and months, although the remaining models may not be compatible with RTX technology, so they will continue with the suffix GTX and it is also possible that they continue to use the Pascal architecture, although none of this has been officially confirmed so we will have to wait to see how it finally unfolds.

This ends our special article dedicated to the new Nvidia RTX graphics cards, remember that you can leave a comment if you have any suggestions or something to add. You can also share the article with your friends on social networks, in this way you help us to spread it so that it can reach more users who need it. What do you think of the arrival of Ray Tracing to the new Nvidia graphics cards? Do you think they should have focused more on improving raster performance?