GeForce RTX 3090 and Stream Processor - video card comparison

GeForce RTX 3090

Stream Processor

Architecture

Ampere (2020−2022)

Ultra-Threaded SE

GPU code name

Ampere GA102

R580

Market segment

Desktop

Release date

September 2020

2006

Interface

PCIe 4.0 x16

PCIe 1.0 x16

Supplementary power connectors

1x 12-pin

1x 6-pin

Boost clock speed

1700 MHz

Core clock speed

1400 MHz

594 MHz

Manufacturing process technology

8 nm

90 nm

Number of transistors

28,300 million

384 million

Pipelines / CUDA cores

10496

Power consumption (TDP)

350 Watt

165 Watt

Texture fill rate

556.0

9.504

Memory type

GDDR6X

GDDR3

Maximum RAM amount

24 GB

1 GB

Memory bus width

384 Bit

256 Bit

Memory clock speed

19500 MHz

1296 MBps

Memory bandwidth

936.2 GB/s

41.47 GB/s

Shared memory

Display Connectors

1x HDMI, 3x DisplayPort

2x DVI, 1x S-Video

HDMI

DirectX

12 Ultimate (12_2)

9.0c (9_3)

CUDA

8.5

OpenCL

2.0

N/A

OpenGL

4.6

2.1 (full) Shader Model6.53.0

Vulkan

1.2

N/A