NVIDIA TU104
TU104
TU104-150-KC-A1
TU104-400A-A1
TU104-410-A1
TU104-450-A1
TU104-850-A1
N18E-G3-A1
Block Diagram
SM Diagram
NVIDIA's TU104 GPU uses the Turing architecture and is made using a 12 nm production process at TSMC. With a die size of 545 mm² and a transistor count of 13,600 million it is a very big chip. TU104 supports DirectX 12 Ultimate (Feature Level 12_2). For GPU compute applications, OpenCL version 3.0 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 3072 shading units, 192 texture mapping units and 64 ROPs. Also included are 384 tensor cores which help improve the speed of machine learning applications. The GPU also contains 48 raytracing acceleration cores.
Further reading:
Turing Architecture Whitepaper 
Graphics Processor
- GPU Name
- TU104
- Codename
- NV164
- Architecture
- Turing
- Foundry
- TSMC
- Process Size
- 12 nm
- Transistors
- 13,600 million
- Density
- 25.0M / mm²
- Die Size
- 545 mm²
- Released
- Aug 13th, 2018
Graphics Features
- DirectX
- 12 Ultimate (12_2)
- OpenGL
- 4.6
- OpenCL
- 3.0
- Vulkan
- 1.2
- CUDA
- 7.5
- Shader Model
- 6.6
- PureVideo HD
- VP10
- VDPAU
- Feature Set J
Render Config
- Shading Units
- 3072
- TMUs
- 192
- ROPs
- 64
- SM Count
- 48
- FP16 Units
- 6144
- FP64 Units
- 96
- INT32 Units
- 3072
- Tensor Cores
- 384
- RT Cores
- 48
- SFUs
- 768
- TPCs
- 24
- GPCs
- 6
- Tex L1 Cache
- 32 KB per SM
- L1 Cache
- 64 KB per SM
- L2 Cache
- 4096 KB
- Max. TDP
- 250 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
- 2016-2021 Pascal
- 2017-2020 Volta
- 2018-2021 Turing
- 2020-2021 Ampere
Graphics cards using the NVIDIA TU104 GPU
| Name | Chip | Memory | Shaders | TMUs | ROPs | Base Clock | Boost Clock | Memory Clock |
|---|---|---|---|---|---|---|---|---|
NVIDIA GeForce RTX 2080
|
TU104-400A-A1 | 8 GB | 2944 | 184 | 64 | 1515 MHz | 1710 MHz | 1750 MHz |
NVIDIA Quadro RTX 5000
|
TU104-875-A1 | 16 GB | 3072 | 192 | 64 | 1620 MHz | 1815 MHz | 1750 MHz |
NVIDIA GeForce RTX 2080 Mobile
|
N18E-G3-A1 | 8 GB | 2944 | 184 | 64 | 1380 MHz | 1590 MHz | 1750 MHz |
NVIDIA Tesla T4
|
TU104-895-A1 | 16 GB | 2560 | 160 | 64 | 585 MHz | 1590 MHz | 1250 MHz |
NVIDIA Quadro RTX 4000
|
TU104-850-A1 | 8 GB | 2304 | 144 | 64 | 1005 MHz | 1545 MHz | 1625 MHz |
NVIDIA GeForce RTX 2080 Max-Q
|
N18E-G3-A1 | 8 GB | 2944 | 184 | 64 | 735 MHz | 1095 MHz | 1500 MHz |
| NVIDIA Quadro RTX 4000 Max-Q | 8 GB | 2560 | 160 | 64 | 780 MHz | 1380 MHz | 1625 MHz | |
| NVIDIA Quadro RTX 4000 Mobile | N19E-Q3-A1 | 8 GB | 2560 | 160 | 64 | 1110 MHz | 1560 MHz | 1750 MHz |
| NVIDIA Quadro RTX 5000 Mobile | N19E-Q5-A1 | 16 GB | 3072 | 192 | 64 | 1365 MHz | 1770 MHz | 1750 MHz |
| NVIDIA Quadro RTX 5000 Max-Q | 16 GB | 3072 | 192 | 64 | 600 MHz | 1350 MHz | 1500 MHz | |
NVIDIA GeForce RTX 2080 SUPER
|
TU104-450-A1 | 8 GB | 3072 | 192 | 64 | 1650 MHz | 1815 MHz | 1937 MHz |
NVIDIA GeForce RTX 2070 SUPER
|
TU104-410-A1 | 8 GB | 2560 | 160 | 64 | 1605 MHz | 1770 MHz | 1750 MHz |
NVIDIA GeForce RTX 2060 TU104
|
TU104-150-KC-A1 | 6 GB | 1920 | 120 | 48 | 1365 MHz | 1680 MHz | 1750 MHz |
| NVIDIA GeForce RTX 2080 SUPER Mobile | N18E-G3 | 8 GB | 3072 | 192 | 64 | 1365 MHz | 1560 MHz | 1750 MHz |
| NVIDIA GeForce RTX 2070 SUPER Mobile | N18E-G2R | 8 GB | 2560 | 160 | 64 | 1140 MHz | 1380 MHz | 1750 MHz |
| NVIDIA GeForce RTX 2070 SUPER Max-Q | 8 GB | 2560 | 160 | 64 | 930 MHz | 1155 MHz | 1375 MHz | |
| NVIDIA GeForce RTX 2080 SUPER Max-Q | N18E-G3R | 8 GB | 3072 | 192 | 64 | 735 MHz | 975 MHz | 1375 MHz |
