GeForce RTX 3080 and Radeon Instinct MI250 - video card comparison

GeForce RTX 3080

Radeon Instinct MI250

Architecture

Ampere (2020−2022)

CDNA 2.0 (2021)

GPU code name

Ampere GA102

Aldebaran

Market segment

Desktop

Release date

September 2020

November 2021

Interface

PCIe 4.0 x16

Supplementary power connectors

1x 12-pin

2x 8-pin

Boost clock speed

1710 MHz

1700 MHz

Core clock speed

1450 MHz

Manufacturing process technology

8 nm

6 nm

Number of transistors

28,300 million

58,200 million

Pipelines / CUDA cores

8704

13312

Power consumption (TDP)

320 Watt

500 Watt

Texture fill rate

465.1

1,414

Memory type

GDDR6X

HBM2e

Maximum RAM amount

10 GB

128 GB

Memory bus width

320 Bit

8192 Bit

Memory clock speed

19000 MHz

3.2 GB/s

Memory bandwidth

760.3 GB/s

3,277 GB/s

Shared memory

Display Connectors

1x HDMI, 3x DisplayPort

No outputs

HDMI

DirectX

12 Ultimate (12_2)

N/A

CUDA

8.5

OpenCL

2.0

3.0

OpenGL

4.6

N/A

Shader Model

6.5

N/A

Vulkan

1.2

N/A