NVIDIA Turing

NVIDIA Turing
Name	Turing
Developer	NVIDIA Corporation
Released	2018
Architecture	Turing
Predecessor	Volta
Successor	Ampere
Process	12 nm
Type	GPU microarchitecture

NVIDIA Turing NVIDIA Turing is a GPU microarchitecture introduced by NVIDIA Corporation in 2018 as a successor to Volta and a precursor to Ampere. It combined programmable shading, dedicated ray tracing hardware, and tensor processing to accelerate graphics and compute workloads across fields like real-time rendering, artificial intelligence, and professional visualization. Major product families based on Turing included workstation and consumer graphics cards deployed in applications spanning Blizzard Entertainment, Electronic Arts, Autodesk, Pixar, and research institutions such as Lawrence Livermore National Laboratory and CERN.

Overview

Turing represented a shift toward hybrid rendering paradigms by integrating RT Cores for bounding volume hierarchy traversal and ray/triangle intersection and Tensor Cores for matrix operations used in deep learning frameworks like TensorFlow and PyTorch. The microarchitecture targeted markets served by companies such as Adobe Systems, Unity Technologies, Epic Games, Siemens, and Dassault Systèmes. Announced alongside partnerships with studios including Weta Digital and engine providers like Unreal Engine and Unity (game engine), Turing aimed to accelerate workflows in film production, architectural visualization, and scientific computing.

Architecture

Turing's hardware design integrated several specialized units: RT Cores, Tensor Cores, and improved CUDA cores, together with advanced memory subsystems and cache hierarchies. The design built on ideas from predecessors employed at organizations like NVIDIA Research and referenced techniques studied at universities including Stanford University, MIT, UC Berkeley, and University of Toronto. The microarchitecture used a refined streaming multiprocessor layout, updated instruction set features, and integer/float mixed-precision pipelines that supported workloads from companies such as Google and Facebook. Fabrication was handled by foundries akin to those servicing TSMC and Samsung Electronics at comparable process nodes.

Products and Implementations

Products based on Turing spanned consumer, prosumer, and professional lines: gaming cards used by Valve Corporation and Riot Games players; workstation GPUs adopted by Industrial Light & Magic and Framestore; and mobile variants for laptop manufacturers like ASUS, Dell, HP, Lenovo, and MSI. Server and accelerator form factors were incorporated into offerings sold to cloud providers such as Amazon Web Services, Google Cloud Platform, Microsoft Azure, and HPC centers at institutions like NASA and Oak Ridge National Laboratory. OEM integrations involved companies including Intel Corporation for platform compatibility and Microsoft for driver and OS support.

Performance and Benchmarks

Benchmarks published by hardware reviewers and research labs compared Turing against predecessors and contemporaries from competitors like AMD and historical products from Intel. Measurements used industry-standard suites and workloads from organizations such as SPEC, 3DMark, Blender Foundation, and academic benchmarks from Stanford DAWN and MLPerf. In ray-tracing scenarios Turing showed significant gains in hybrid-rendering throughput, while Tensor Core acceleration improved mixed-precision training and inference times for models used by OpenAI and research groups at Carnegie Mellon University. Comparative analysis influenced purchasing decisions at enterprises including Walmart Labs and Goldman Sachs for accelerated compute tasks.

Features and Technologies

Key technologies included real-time ray tracing support via RT Cores, deep learning acceleration via Tensor Cores, and support for APIs and frameworks such as DirectX 12, Vulkan, OpenGL, CUDA, OptiX, and integrations with engines like Unreal Engine and middleware from NVIDIA partners. Other features encompassed memory compression, variable rate shading, and programmable mesh shading concepts that intersected with research from institutions like ETH Zurich and University of Cambridge. Software ecosystems and developer tools tied into platforms from Microsoft Visual Studio, GitHub, and middleware from Autodesk and SideFX.

Market Reception and Impact

Turing's market reception combined acclaim for its forward-looking hardware with critique about pricing and power consumption noted by reviewers at outlets including AnandTech, Tom's Hardware, TechRadar, PC Gamer, and The Verge. It influenced adoption of hardware-accelerated ray tracing in games from studios like Square Enix and Bethesda Softworks and drove feature support in engines such as Frostbite and CryEngine. Enterprises across sectors—media companies like Netflix for encoding workflows, automotive suppliers such as Bosch and Continental for simulation, and healthcare research groups at Johns Hopkins University—reported productivity gains. The architecture also affected competitive roadmaps at firms including AMD and reinforced NVIDIA's positioning with customers like Hewlett Packard Enterprise and cloud providers.

Category:Graphics processing units