Radeon VII and Tesla M40 - video card comparison

Passmark - benchmark score Radeon VII vs Tesla M40

Radeon VII

17191

Tesla M40

10687

Radeon VII

Tesla M40

Architecture

Vega (2017−2021)

Maxwell 2.0 (2015−2019)

GPU code name

Vega 20

GM200

Market segment

Desktop

Workstation

Release date

February 2019

November 2015

Interface

PCIe 3.0 x16

Supplementary power connectors

2x 8-pin

8-pin EPS

Boost clock speed

1750 MHz

1112 MHz

Core clock speed

1400 MHz

948 MHz

Floating-point performance

6,844 gflops

Manufacturing process technology

7 nm

28 nm

Number of transistors

13,230 million

8,000 million

Pipelines / CUDA cores

3840

3072

Power consumption (TDP)

295 Watt

250 Watt

Texture fill rate

420.0

213.5

Memory type

HBM2

GDDR5

Maximum RAM amount

16 GB

12 GB

Memory bus width

4096 Bit

384 Bit

Memory clock speed

1000 MHz

6 GB/s

Memory bandwidth

1024 GB/s

288.4 GB/s

Display Connectors

1x HDMI, 3x DisplayPort

No outputs

HDMI

DirectX

12 (12_1)

CUDA

5.2

OpenCL

2.0

3.0

OpenGL

4.6

Shader Model

6.4

6.7

Vulkan

1.2.131

1.3