Turing (microarchitecture): Difference between revisions
→Tensor core: ref |
No edit summary |
||
| Line 21: | Line 21: | ||
New features in Turing: |
New features in Turing: |
||
* [[CUDA]] Compute Capability (traditional rasterized shaders) |
* [[CUDA]] Compute Capability ?.?(traditional rasterized shaders) |
||
* Raytracing (RT) core |
* Raytracing (RT) core |
||
* Tensor (AI) core |
* Tensor (AI) core |
||
Revision as of 15:35, 21 August 2018
| Release date | 20 September 2018 |
|---|---|
| Fabrication process | 12 nm |
| History | |
| Predecessor | Pascal (consumer), Volta (professional) |
Turing is the codename for a GPU microarchitecture developed by Nvidia, as the successor to the Pascal architecture. It is named after Alan Turing, a prominent computer scientist. The architecture was first introduced in August 2018 at Gamescom, along with Quadro and GeForce products based on it.[1] The architecture introduces the first consumer products capable of real-time raytracing, which has been a longstanding goal of the computer graphics industry. Key elements include a dedicated artificial intelligence processor ("Tensor core") and dedicated raytracing processor. Turing leverages Microsoft's DXR, OptiX and Vulkan for access to raytracing.
Details
The Turing microarchitecture principally features real-time raytracing.[2] This is accelerated by the use of new RT (raytracing) cores, which are designed to process quadtrees and spherical hierarchies, and speed up collision tests with individual triangles.
New features in Turing:
- CUDA Compute Capability ?.?(traditional rasterized shaders)
- Raytracing (RT) core
- Tensor (AI) core
- Memory controller with GDDR6 support
- GPU Boost 4
- NVLink Bridge with VRAM stacking pooling memory from multiple cards
- VirtualLink VR
Raytracing
The ray-tracing performed by the RT cores can be used to produce reflections, refractions and shadows, replacing traditional raster techniques such as cube maps and depth maps. Instead of replacing rasterization entirely, however, the information gathered from ray-tracing can be used to augment the shading with information that is much more photo-realistic, especially in regards to off-camera action. Nvidia said the raytracing performance increased about 8 times over the previous consumer architecture, Pascal.
Tensor core
Ray-tracing of a final image is further accelerated by the included Tensor cores, which are used to fill in the blanks in a partially rendered image, a technique known as de-noising. The Tensor core performs the result of deep learning on supercomputers to codify how to, for example, increase the resolution of images. In the Tensor core's primary usage, a problem to be solved is analyzed on a supercomputer, which is taught by example what results are desired, and the supercomputer determines a method to use to achieve those results which is then done with the consumer's Tensor core. These methods are delivered "over the air" to consumers.[3]
Development
Turing's development platform is called RTX. RTX leverages Microsoft's DXR, OptiX and Vulkan for access to raytracing.[4] It includes access to AI-accelerated features through NGX.
Products using Turing
- GeForce RTX 2080 Ti
- GeForce RTX 2080
- GeForce RTX 2070
- Quadro RTX 8000
- Quadro RTX 6000
- Quadro RTX 5000
See also
References
- ^ "Nvidia reveals $800 GeForce RTX 2080 at Gamescom 2018". CNet.
- ^ "Nvidia announces RTX 2000 GPU series with '6 times more performance' and ray-tracing". The Verge. Retrieved 20 August 2018.
- ^ "#BeForTheGame". Twitch.tv.
- ^ "NVIDIA RTX™ platform". Nvidia.