What is NVIDIA® Turing™ Architecture?

If you follow the latest graphical processing capabilities, you'll know that Turing™ is the name of NVIDIA's® latest graphics processing unit (GPU) architecture. This NVIDIA architecture is named after famous mathematician and code breaker Alan Turing OBE FRS, one of the founders of theoretical computer science. You will find the NVIDIA Turing architecture in many of NVIDIA's currently available products.

The NVIDIA Turing architecture is notable because it introduced real-time ray tracing to the consumer market. In addition, this advanced rendering technique has brought professional cinematic quality image generation to the home gaming market.

History of the NVIDIA Turing Architecture

Since the formation of NVIDIA in 1993, it has consistently developed and released innovative graphical processing units.

• In 2006, the Tesla architecture employed 90 nm semiconductor fabrication to produce the first integrated solution based on stream processors with load balancing capability for efficient image generation.
• In 2010, the Fermi architecture employed 40 nm semiconductor fabrication to evolve the Tesla architecture into a more efficient processing unit. This architecture brought significant processing throughput benefits and the introduction of a polymorph engine for scalable geometric calculations.
• In 2012, the Kepler architecture employed 28 nm semiconductor fabrication to produce a more energy-efficient development of the Fermi architecture using lower clock speeds. This efficiency improvement reduced power consumption and thermal issues associated with internal heat generation without compromising performance.
• In 2014, the Maxwell architecture advanced the Fermi semiconductor fabrication to produce a more energy-efficient development similar to the Kepler architecture but with lower complexity. This feature created a processing core with smaller physical dimensions and lower power consumption suited to smaller mobile devices.
• In 2016, the Pascal architecture employed 16 nm semiconductor fabrication to evolve the Maxwell architecture to increase performance by adding more streaming processing cores and improving memory access mechanisms.
• In 2018, the Turing architecture represented a significant revolution of the Pascal architecture to add novel functionality. It employed 12 nm semiconductor fabrication to pack in multiple specialised processing cores for streaming, ray tracing, and artificial intelligence-based processing functions. This significantly increased processing capabilities beyond simple performance improvements. As a result, the available capabilities revolutionise the gaming experience for the consumer market.

Key Features of the NVIDIA Turing Architecture

The basic philosophy behind the NVIDIA Turing architecture is leveraging parallel processing to generate high-quality three-dimensional graphics for computationally intensive gaming applications.

The top-of-the-range Turing TU102 GPU chipset includes 6 Graphics Processing Clusters (GPC). These GPCs have a dedicated raster engine and 6 Texture Processing Clusters (TPC), totalling 36 TPCs. Each TPC, in turn, includes 2 Streaming Multiprocessors (SM). This gives a total of 72 SMs. Each SM contains 64 Compute Unified Device Architecture (CUDA™) Cores, 8 Tensor™ Cores, a 256 KB register file, 4 texture units, and 96 KB of configurable memory.

The CUDA Cores perform the image processing using parallel computational elements. The Tensor Cores perform the artificial intelligence-based processing based on neural networks. Therefore, the TU102 GPU chipset includes a total of 4,608 CUDA Cores and 576 Tensor Cores.

Hardware-based accelerators combined with an innovative hybrid rendering technique improve processing throughput over existing graphics processing capabilities. Specialised processing cores fuse graphical functionality, including real-time ray tracing and rasterisation, with artificial intelligence-based processing and simulation functions. These features have the power and ability to generate cinematic-quality interactive experiences with seamless fluidity for interactive gaming and three-dimensional modelling applications.

The key innovation is the integral real-time ray tracing core. This development eliminates the requirement for software emulation-based ray tracing techniques that are prohibitively expensive for most consumer applications. Instead, the real-time ray-traced rendering creates photorealistic objects and environments with physically accurate shadowing effects, reflections, and refractions for everyday applications.

The NVIDIA Turing architecture also includes more advanced object shading capabilities that can handle greater geometric complexity with improved image quality than previous architectures.

Also, the artificial intelligence-based processing capability enables gaming applications to incorporate real-time deep learning techniques to enhance the user experience. In particular, these techniques can create crisp, clear, and lifelike images that allow realistic, immersive gaming experiences.

The NVIDIA Turing architecture also offers connectivity improvements. In addition to supporting the second-generation NVIDIA NVLink™, it also supports both USB Type-C™ and VirtualLink™ interfaces used by the next generation of virtual reality (VR) headsets.

List of NVIDIA Products that use the Turing Architecture

NVIDIA has implemented its Turing architecture in the TU102, TU104, TU106, TU116, and TU117 chipsets. You can find the NVIDIA Turing architecture in a range of popular graphical processing products that use these chipsets, including:

• The GeForce® Series 16 products (including GTX 1650 and GTX 1660 ranges) aim to satisfy high-performance graphical processing for the entry-level to mid-range GPU market.
• The GeForce Series 20 products (including RTX 2060, RTX 2070, and RTX 2080 ranges) aim to satisfy high-performance graphical processing for the mid-range to high-end GPU market.
• The NVIDIA Quadro® graphics cards (including RTX 3000, 4000, 5000, 6000, 8000, T400, T600, and T1000) aim to satisfy high-performance graphical processing for workstations used for professional design applications.
• The NVIDIA T4 GPU accelerator general-purpose graphics processing unit (GPGPU) aims to satisfy graphical processing needs for high-performance workstations.

HP Products With NVIDIA Turing Architecture

1. Mid-Range Laptop - HP Pavilion Gaming Laptop

Shop HP Pavilion Gaming Laptop 15-dk2060na Now!

The HP Pavilion Gaming Laptop 15-dk2060na - NVIDIA GeForce GTX 1650 is an excellent example of a mid-range laptop that incorporates the NVIDIA Turing architecture into its graphics processor. Designed for mobile gaming, this compact laptop with an 11th Generation Intel® Core™ i5 processor combines the processing power and graphical capability to handle image-intensive gaming with smooth gameplay visuals. In addition, the NVIDIA GeForce GTX 1650 graphics card with 4GB of dedicated GDDR6 memory delivers fast, smooth, power-efficient performance.

This thin and powerful HP Pavilion Gaming Laptop offers users high-grade graphics and processing power ideal for gaming and multitasking. In addition, it incorporates thermal cooling improvements for overall performance and stability for sustained operations. These features allow users to immerse themselves in gaming applications.

2. High-End Laptop - HP ZBook Create G7

Shop HP ZBook Create G7 DreamColor Mobile Workstation Now!

The HP ZBook Create G7 15.6" 4K DreamColor Mobile Workstation with i9 & NVIDIA GeForce RTX 2070 is an excellent example of a high-end laptop that incorporates the NVIDIA Turing architecture. Designed for creating and gaming on the go, this small 15" mobile workstation with a 10th Generation Intel Core i9 processor combines the processing power and graphical capability to meet challenging user needs. In addition, the NVIDIA GeForce RTX 2070 graphics card with 8GB of dedicated GDDR6 memory supports intensive image processing tasks, including rendering and video editing.

The HP ZBook Create G7is HP's smallest lightweight laptop designed for gaming and content creation workflows. Vapour chamber cooling techniques and liquid-crystal polymer thermal management support sustained use for intensive gaming applications. Its high performance for gamers is down to the combination of the NVIDIA GeForce graphics with HP's Z Turbo Drive high-speed storage.

3. Mid-Range Desktop Workstation - HP Z2 G8

Shop HP Z2 G8 Small Form Factor Workstation Now!

The HP Z2 G8 Small Form Factor Workstation with i7 (8 cores) & NVIDIA Quadro T1000 4GB Graphics is an excellent example of a mid-range desktop workstation that incorporates the NVIDIA Turing architecture into its graphics processor. HP® has designed this desktop to provide high-performance graphical processing in a compact workstation intended for professional computer-aided design applications. In addition, the NVIDIA Quadro T1000 graphics card with 4GB of dedicated GDDR6 memory provides high-end visualisation and seamless rendering for image processing tasks.

This adaptable compact desktop can cope with the real-world productivity demands of content-creation workflows, including the Autodesk® and SOLIDWORKS® applications. In addition, tool-less access allows simple configuration upgrades to maintain through-life performance under changing requirements.

4. High-End Desktop Workstation - HP Z1 G6

Shop HP Z1 G6 with i9-10900 & NVIDIA® GeForce® RTX 2080 Now!

The HP Z1 G6 with i9-10900 & NVIDIA GeForce RTX 2080 (8GB) is an excellent example of a high-end desktop workstation incorporating the NVIDIA Turing architecture into its graphics processor. HP® has designed this desktop to deliver high-performance graphical processing intended for professional computer-aided design applications and VR content creation. The NVIDIA GeForce RTX 2080 graphics card with 8GB of dedicated GDDR6 memory provides high-end performance for intensive image processing tasks, including rendering and video editing.

This flexible, configurable desktop is tested and certified for use with a range of professional software packages, including AutoCAD, SOLIDWORKS, and Revit®, guaranteeing seamless, reliable workflow performance.

Summary

Complex image rendering and processing tasks require high-performance graphical computing capabilities. Therefore, choosing the right GPU for your laptop or desktop is essential to get the best performance for image-intensive applications ranging from content creation to immersive gaming.

GPUs based on the NVIDIA Turing architecture can support complex real-time ray tracing and rasterisation using hardware-based accelerators that maximise performance. In addition, artificial intelligence-based processing support makes such GPUs ideal for the most challenging creation, simulation and virtual reality applications, and complex three-dimensional modelling applications.