NX-SoC and Quadro RTX 8000 - video card comparison
NX-SoC
Quadro RTX 8000
Architecture
Maxwell 2.0 (2015−2019)
Turing (2018−2021)
GPU code name
GM20B
TU102
Market segment
Desktop
Workstation
Release date
March 2017
August 2018
Interface
IGP
PCIe 3.0 x16
Supplementary power connectors
None
1x 6-pin + 1x 8-pin
Boost clock speed
768 MHz
1770 MHz
Core clock speed
384 MHz
1395 MHz
Floating-point performance
393.2 gflops
-
Manufacturing process technology
20 nm
12 nm
Number of transistors
2,000 million
18,600 million
Pipelines / CUDA cores
256
4608
Power consumption (TDP)
15 Watt
260 Watt
Texture fill rate
12.29
509.8
Memory type
DDR4
GDDR6
Maximum RAM amount
4 GB
48 GB
Memory bus width
64 Bit
384 Bit
Memory clock speed
3200 MHz
14000 MHz
Memory bandwidth
25.6 GB/s
672.0 GB/s
Display Connectors
No outputs
4x DisplayPort, 1x USB Type-C
DirectX
12 (12_1)
12 Ultimate (12_1)
CUDA
5.3
7.5
OpenCL
1.2
2.0
OpenGL
4.6
4.6
Shader Model
6.4
6.5
Vulkan
1.1.126
1.2.131