Latte and Jetson Nano GPU - video card comparison

Latte

Jetson Nano GPU

Architecture

TeraScale 2 (2009−2015)

Maxwell 2.0 (2015−2019)

GPU code name

Latte

GM20B

Market segment

Desktop

Desktop

Release date

-

March 2019

Interface

-

IGP

Boost clock speed

-

921 MHz

Manufacturing process technology

40 nm

20 nm

Number of transistors

880 million

2,000 million

Pipelines / CUDA cores

160

128

Power consumption (TDP)

-

10 Watt

Texture fill rate

-

14.74

Memory type

-

System Shared

Maximum RAM amount

-

System Shared

Memory bus width

-

System Shared

Memory clock speed

-

System Shared

Display Connectors

-

No outputs

DirectX

N/A

12 (12_1)

CUDA

-

5.3

OpenCL

N/A

1.2

OpenGL

N/A

4.6

Shader Model

5.0

6.4

Vulkan

N/A

1.1