NVIDIA TU102 Graphics Processing Unit (GPU)

Overview
Manufacturer
NVIDIA
Series
Quadro Turing
Release Date
December 31st, 2018
Graphics Processing Unit
GPU Model
TU102-875 (TU102)
Architecture
Turing
Fabrication Process
12 nm FinFET
Die Size
754 mm2
Transistors Count
18.6B
Transistors Density
24.7M TRAN/mm2
CUDAs
4608
TMUs
288
ROPs
96
Clocks
Base Clock
1440 MHz
Boost Clock
1770 MHz
Memory Clock
1625 MHz
Effective Memory Clock
13000 Mbps
Memory Configuration
Memory Size
24576 MB
Memory Type
GDDR6
Memory Bus Width
384-bit
Memory Bandwidth
624.0 GB/s
Physical
Interface
PCI-Express 3.0 x16
Power Connectors
2 × 8-pin
Thermal Design Power
295 W
Recommended PSU
600 W
SLI Support
2-way
Display Outputs
DisplayPort
4 ×
API Support
DirectX
12 Ultimate
Mantle
-
Vulkan
1.0
OpenGL
4.5
OpenCL
2.1
Shader Model
-

Performance
Pixel Fillrate
138.2 GPixels/s
Texture Fillrate
414.7 GTexel/s
SPFP Performance
16.3 TFLOPS
Performance per W
55.3 GFLOPS/W
Performance per mm2
21.6 GFLOPS/mm2


  Model GPU Clocks Memory
NVIDIA Quardro RTX 6000 TU102-875 1440 / 1625 MHz 24 GB GDDR6
NVIDIA Quardro RTX 6000 Mobile TU102 1616 / 1750 MHz 24 GB GDDR6
NVIDIA Quardro RTX 6000 Passive TU102-875 1305 / 1625 MHz 24 GB GDDR6
NVIDIA Quardro RTX 8000 TU102-875 1395 / 1750 MHz 48 GB GDDR6
NVIDIA Quardro RTX 8000 Passive TU102-875 1230 / 1750 MHz 48 GB GDDR6
NVIDIA Tesla RTX T10-8 TU102 1300 / 1750 MHz 24 GB GDDR6
NVIDIA TITAN RTX TU102-400 1350 / 1750 MHz 24 GB GDDR6
NVIDIA GeForce RTX 2080 Ti TU102-300/A 1350 / 1750 MHz 11 GB GDDR6
NVIDIA GeForce RTX 2080 Ti Cyberpunk 2077 Edition TU102-300A 1350 / 1750 MHz 11 GB GDDR6
NVIDIA GeForce RTX 2080 Ti Founders Edition (V)
4352 Cores
TU102 1350 / 1750 MHz 11 GB GDDR6