NVIDIA is the leader in the graphics card market and has many forward-looking technologies to help keep it at the forefront. In this article, we will tell you about the most important and significant technologies of this company.
In this article, we will not consider old technologies in adapters until 2010. Also, technologies in very highly specialized products will not be considered here. If you want, you can look at the technologies on the NVIDIA website.
You may also be interested in materials on our website about technologies from other chip manufacturers (links will be below in the article), markings and lines of NVIDIA video cards, integrated Intel graphics, interesting facts about video cards.
Screen and display technologies
NVIDIA 3D Vision
This is a kind of ecosystem for 3D from NVIDIA. It includes support from video cards, special monitors, special glasses and a lot of content: games, movies, photos and videos from YouTube.
Monitors / projectors that support this technology have a large 120 Hz hertz.
This technology must be supported by glasses and a 3D monitor. It allows the image to be brighter than normal active shutter 3D glasses. It can also reduce ghosting.
With this technology, you can create a large panorama from multiple monitors / projectors and it will work like a large workspace. You can also combine 3D monitors with 3D VISION support.
It allows you to connect your PC or laptop to a 3D TV and play on it. Both active and passive 3D creation technologies are supported.
It is a technology for monitors and video cards that allows them to sync and frame at the same rate. In this way, a smooth image is achieved without tearing, jerking or lag. The main requirement is for a monitor, because it must be certified for this technology.
G-Sync synchronizes the frequency of the monitor and the FPS, which is issued by the video card. This technology, unlike vertical sync, does not create a delay in displaying the image on the screen.
It is a frame capture analysis technology. It allows you to more accurately measure the number of FPS compared to FRAPS. It is able to detect dropped frames, dwarf frames, micro-jerks, and other problems. This allows you to more accurately track the performance of the video card.
This technology is used in laptops. It consists in the fact that it allows you to switch from the graphics integrated into the processor to a separate discrete chip. Discrete graphics have a performance advantage, while built-in graphics have an advantage in energy efficiency. This allows us to partially eliminate the dilemma: “What is better when choosing a video card for a laptop: discrete graphics or integrated ?, because this technology allows you to take advantage of both types.
Unlike other similar technologies, it is possible to customize the use of graphics in individual applications. This means that now the selected graphics depends not only on the chosen power scheme, but on the requirements of the program. It is also possible to manually configure the selected graphics solution for different applications.
Allows you to increase the battery life of the laptop in games by dynamically changing performance.
This means that if the game scene is currently easy to render, the video card will slow down performance. The FPS doesn’t drop, it just doesn’t jump at such light moments. It is possible to adjust the acceptable frame rate in different games. Thus, it is possible to play longer on battery power, which is very pleasant. Also, this technology interacts with the previous one.
It is available starting with the 800M series and the GeForce GTX 850M.
This technology is designed to dynamically change the frequency and voltage of a video card depending on temperature and load.
Only professional benchmarks load video cards to the full and “warm them up to critical temperatures. But even the most intense games are not capable of this, and in the game we have a margin of temperature and performance. GPU Boost allows you to take advantage of this headroom by increasing the frequency and voltage.
This technology has several versions. The advantage of the second is that it gives the user more control. In this way, the minimum acceptable temperature can be changed as well as the operating level of the fans. Also, version 2.0 allows you to further increase the frequency until the video card warms up. This is very reminiscent of the story with Intel Turbo Boost 2.0, which you can read about here.
With the introduction of the Pascal generation, this technology has version 3.0. Unlike 2.0, where the increase in frequency is linear and directly proportional to the increase in voltage, there are several frequency points for each voltage point. In the special software for overclocking, it is already possible to adjust the frequency for different voltages. Thus, at the same voltage, this technology will squeeze out a higher frequency (up to the temperature limit). It turns out that the new version of GPU Boost makes it possible to fine-tune the frequency and get the maximum stable frequency under these conditions. The picture below is clickable.
Connections and connections
This technology allows multiple graphics cards to be combined on a dedicated certified motherboard for increased graphics performance. This technology also requires video adapters supporting this technology and a special SLI bridge. You can do without it, but then the performance will be lower. By the way, AMD has a similar AMD CrossfireX technology.
This technology has special modes of division of work between video cards in multi-screen mode and when using PhysX, as well as several modes of operation in graphics applications, which can be found at the links: multi-screen mode, PhysX, modes of operation.
SLI also has a bunch of disadvantages:
- the amount of video memory is not something that does not add up, so also when using different cards, the bundle uses the smaller volume of all cards
- difficulties arise when using cards of different performance
- creation of a bundle from cards from different manufacturers (NVIDIA and AMD) is not supported
- if the game does not support this technology, there may be no performance gain
This is the new standard for interconnection between CPU and GPU, and between multiple GPUs. It is 5 to 12 times faster than the current PCIe Gen3. This will increase performance in bandwidth-critical applications.
This is a standard for connecting a video card in a laptop using a mobile PCI Express module, when there is a possibility of its subsequent replacement. So if you want to replace a discrete video card in a laptop, you will not need to buy a whole new system. This technology has been worked on not only by NVIDA, but also by various laptop manufacturers.
It is a technology for parallel computing using the power of the graphics chip. This provides a powerful computing tool for economic forecasts, weather forecasting, mathematical calculations, physics calculations, etc. A GPU will do these tasks better than a CPU due to the large number of compute units.
To do this, you need to specially optimize the application for hardware acceleration CUDA.
This technology debuted with the new generation of Pascal. It allows you to take specially prepared screenshots with many settings, this process is similar to taking a photo. Here you can change the screenshot resolution up to 32x compared to the resolution of your monitor, there are filters, panoramic shooting, 360 ° and much more.
This technology must be supported at the game level. The first such games will be The Division, The Witness, Lawbreakers, The Witcher 3: Wild Hunt, Paragon, No Man’s Sky and Unreal Tournament.
It is a technology aimed at remote hardware acceleration of graphics and computing. This means that the operations you need are performed on the GRID servers and only the result is sent to you. Applications are launched remotely, and a kind of translation goes to the computer, this is called application virtualization. By the way, you can do this with games too.
There are several versions for different types of users with different capacity limits. This technology also implements data protection.
This is an extension for overclocking RAM modules. So on such modules there is the SLI Ready marking, although “overclocking RAM has a very distant relation to this. This RAM has a place for storing SDP profiles, into which you can load optimal overclocking parameters from other users. This RAM will only work on a motherboard with a dedicated chipset.
There are two versions:
- EPP DDR2
- EPP 2.0 DDR3
AMD has a similar technology with their AMP and Intel with their XMP about which you can read the articles on the links.
This is a toolkit for game developers that aims to create VR-enabled games.
This is a rendering method in which the picture is oval rather than rectangular as usual, i.e. where the lens system goes, the video card will not render blind areas. Thanks to the features of the Maxwell architecture, VR games will be able to experience a performance boost. This is achieved by using 9 viewpoints in a single pass. This technology also works well on Pascal graphics cards. The first cards of the new generation were presented on May 6.
Simultaneous Multi Projection
And this is the technology that was developed for Pascal. This technology already uses up to 16 viewpoints and can render images for multi-monitor solutions, curved displays, VR, taking into account the angles of the displays and their curvature, which allows you to create a more natural image. For VR, this technology is interesting due to the fact that using 16 viewpoints, you can calculate 2 frames for a 3D display in one pass (Single Pass Stereo). And also the Lens Matched Shading technology allows you not to render extra pixels for a rounded image of VR glasses.
Simultaneous Multi Projection works a little more efficiently than Multi-Resolution Shading, but you should not confuse them, since they have different applications and they do not replace each other.
This technology reduces latency in response to head movement and improves fluidity. This is possible due to the fact that Asynchronous Time Warp edits the frame when the head is turned and does not need to be redrawn.
Allows to use SLI-bundle of 2 video cards for sequential rendering separately for each eye.
There is a plug-and-play connection for VR headsets, making it easy to use. This is how glasses are recognized by Windows as a special device and the desktop is not displayed on it, which happens quite often on other systems.
Front Buffer Rendering
The GPU can draw frames directly to the front buffer, which reduces latency.
This technology was also announced together in Pascal. It allows you to create more realistic surround sound by calculating the reflections of sound waves from objects in the virtual world in real time, which enhances the effect of presence. For this, the NVIDIAOptiX ray tracing engine is used.
Technologies for professional cards Quadro and Tesla were not considered here, because there are a lot of them (you can write a separate article about them, if you ask) and they are almost never interesting for ordinary users.
If you find some mistake or lack of some technologies, then first we advise you to look into the section “Introduction, and then you can write your complaints in the comments, because they are always open for you!
We hope that our article helped you to understand a huge number of NVIDIA technologies, as well as help you in the discussion and choosing a new video card.