P102-100 and GeForce RTX 3070 - video card comparison

P102-100

GeForce RTX 3070

Architecture

Pascal (2016−2021)

Ampere (2020−2022)

GPU code name

GP102

Ampere GA104

Market segment

Desktop

Desktop

Release date

February 2018

September 2020

Interface

PCIe 3.0 x4

PCIe 4.0 x16

Supplementary power connectors

2x 8-pin

1x 12-pin

Boost clock speed

1683 MHz

1725 MHz

Core clock speed

1582 MHz

1500 MHz

Floating-point performance

10,125 gflops

-

Manufacturing process technology

16 nm

8 nm

Number of transistors

11,800 million

17,400 million

Pipelines / CUDA cores

3200

5888

Power consumption (TDP)

250 Watt

220 Watt

Texture fill rate

336.6

317.4

Memory type

GDDR5X

GDDR6

Maximum RAM amount

5 GB

8 GB

Memory bus width

320 Bit

256 Bit

Memory clock speed

11008 MHz

14000 MHz

Memory bandwidth

440.3 GB/s

448.0 GB/s

Shared memory

-

-

Display Connectors

No outputs

1x HDMI, 3x DisplayPort

HDMI

-

+

DirectX

12 (12_1)

12 Ultimate (12_2)

CUDA

6.1

8.5

OpenCL

1.2

2.0

OpenGL

4.6

4.6

Shader Model

6.4

6.5

Vulkan

1.2.131

1.2