NVIDIA TU102
TU102
TU102-300A-K1-A1
TU102-300-K2-A1
TU102-300-K3-A1
TU102-875-A1
Die Shot
Block Diagram
SM Diagram
NVIDIA's TU102 GPU uses the Turing architecture and is made using a 12 nm production process at TSMC. With a die size of 754 mm² and a transistor count of 18,600 million it is a very big chip. TU102 supports DirectX 12 Ultimate (Feature Level 12_2). For GPU compute applications, OpenCL version 1.2 and CUDA 7.5 can be used. Additionally, the DirectX 12 Ultimate capability guarantees support for hardware-raytracing, variable-rate shading and more, in upcoming video games. It features 4608 shading units, 288 texture mapping units and 96 ROPs. Also included are 576 tensor cores which help improve the speed of machine learning applications. The GPU also contains 72 raytracing acceleration cores.
Further reading:
Turing Architecture Whitepaper 
Graphics Processor
- GPU Name
- TU102
- Codename
- NV162
- Architecture
- Turing
- Foundry
- TSMC
- Process Size
- 12 nm
- Transistors
- 18,600 million
- Density
- 24.7M / mm²
- Die Size
- 754 mm²
- Released
- Aug 13th, 2018
Graphics Features
- DirectX
- 12 Ultimate (12_2)
- OpenGL
- 4.6
- OpenCL
- 1.2
- Vulkan
- 1.2
- CUDA
- 7.5
- Shader Model
- 6.5
- PureVideo HD
- VP10
- VDPAU
- Feature Set J
Render Config
- Shading Units
- 4608
- TMUs
- 288
- ROPs
- 96
- SM Count
- 72
- FP16 Units
- 9216
- FP64 Units
- 144
- INT32 Units
- 4608
- Tensor Cores
- 576
- RT Cores
- 72
- SFUs
- 1152
- TPCs
- 36
- GPCs
- 6
- Tex L1 Cache
- 32 KB per SM
- L1 Cache
- 64 KB per SM
- L2 Cache
- 6144 KB
- Max. TDP
- 280 W
All Turing GPUs
NVIDIA GPU Architecture History
- 1998-2000 Fahrenheit
- 1999-2005 Celsius
- 2001-2003 Kelvin
- 2003-2005 Rankine
- 2003-2013 Curie
- 2006-2010 Tesla
- 2007-2013 Tesla 2.0
- 2010-2016 Fermi
- 2010-2013 VLIW Vec4
- 2010-2016 Fermi 2.0
- 2012-2018 Kepler
- 2013-2015 Kepler 2.0
- 2014-2017 Maxwell
- 2014-2019 Maxwell 2.0
- 2016-2020 Pascal
- 2017-2020 Volta
- 2018-2020 Turing
- 2020-2021 Ampere
Graphics cards using the NVIDIA TU102 GPU
| Name | Chip | Memory | Shaders | TMUs | ROPs | Base Clock | Boost Clock | Memory Clock |
|---|---|---|---|---|---|---|---|---|
NVIDIA GeForce RTX 2080 Ti
|
TU102-300A-K1-A1 | 11 GB | 4352 | 272 | 88 | 1350 MHz | 1545 MHz | 1750 MHz |
NVIDIA Quadro RTX 8000
|
TU102-875-A1 | 48 GB | 4608 | 288 | 96 | 1395 MHz | 1770 MHz | 1750 MHz |
NVIDIA Quadro RTX 6000
|
TU102-875-A1 | 24 GB | 4608 | 288 | 96 | 1440 MHz | 1770 MHz | 1750 MHz |
NVIDIA TITAN RTX
|
TU102-400-A1 | 24 GB | 4608 | 288 | 96 | 1350 MHz | 1770 MHz | 1750 MHz |
NVIDIA Quadro RTX 6000 Passive
|
TU102-875-A1 | 24 GB | 4608 | 288 | 96 | 1305 MHz | 1560 MHz | 1750 MHz |
NVIDIA Quadro RTX 8000 Passive
|
TU102-875-A1 | 48 GB | 4608 | 288 | 96 | 1230 MHz | 1620 MHz | 1750 MHz |
| NVIDIA Quadro RTX 6000 Mobile | 24 GB | 4608 | 288 | 96 | 1275 MHz | 1455 MHz | 1750 MHz | |
| NVIDIA GRID RTX T10-4 | TU102-875-A1 | 4 GB | 4608 | 288 | 96 | 1065 MHz | 1395 MHz | 1750 MHz |
| NVIDIA GRID RTX T10-8 | TU102-875-A1 | 8 GB | 4608 | 288 | 96 | 1065 MHz | 1395 MHz | 1750 MHz |
| NVIDIA GRID RTX T10-16 | TU102-875-A1 | 16 GB | 4608 | 288 | 96 | 1065 MHz | 1395 MHz | 1575 MHz |