Tesla M2090 and Tesla A100 - video card comparison

Tesla M2090

Tesla A100

Architecture

Fermi 2.0 (2010−2014)

Ampere (2020−2022)

GPU code name

GF110

GA100

Market segment

Workstation

Workstation

Release date

July 2011

-

Interface

PCIe 2.0 x16

PCIe 4.0 x16

Supplementary power connectors

1x 6-pin + 1x 8-pin

None

Core clock speed

651 MHz

1110 MHz

Floating-point performance

1,331.2 gflops

-

Manufacturing process technology

40 nm

7 nm

Number of transistors

3,000 million

38,000 million

Pipelines / CUDA cores

512

7936

Power consumption (TDP)

250 Watt

260 Watt

Texture fill rate

41.66

550.6

Memory type

GDDR5

HBM2e

Maximum RAM amount

6 GB

32 GB

Memory bus width

384 Bit

4096 Bit

Memory clock speed

3696 MHz

2400 MHz

Memory bandwidth

177.4 GB/s

1,229 GB/s

Display Connectors

No outputs

No outputs

DirectX

12 (11_0)

12 (12_1)

CUDA

2.0

-

OpenCL

1.1

-

OpenGL

4.6

4.6

Shader Model

5.1

-

Vulkan

N/A

-