pchardware.org
Categories

NVIDIA · 2018

NVIDIA S2070 GPU Computing Server

The NVIDIA S2070 GPU Computing Server is a graphics card based on the Turing architecture, released in October 2018.

The NVIDIA S2070 GPU Computing Server is a graphics card based on the Turing architecture, featuring the TU106-400A-A1 GPU chip manufactured on a 12 nm process. It includes 2304 shading units, 36 ray tracing cores, and 288 tensor cores. The card operates at a base clock of 1410 MHz and a boost clock of 1620 MHz. It is equipped with 8 GB of GDDR6 memory on a 256-bit bus, providing a memory bandwidth of 448.0 GB/s. The board power (TDP) is 175 W, requiring a single 8-pin power connector. It interfaces via PCIe 3.0 x16 and offers display outputs including 1x HDMI 2.0b, 3x DisplayPort 1.4a, and 1x USB Type-C. The card supports DirectX 12 Ultimate, Vulkan 1.3, and OpenGL 4.6. It was released on October 17, 2018 with a launch MSRP of 499 USD.

Specifications

Architecture unverified Turing
GPU chip unverified TU106-400A-A1
Process node unverified 12 nm
Shading units unverified 2304
Ray tracing cores unverified 36
Tensor / AI cores unverified 288
Base clock unverified 1410 MHz
Boost clock unverified 1620 MHz
Memory size unverified 8 GB
Memory type unverified GDDR6
Memory bus unverified 256 bit
Memory bandwidth unverified 448.0 GB/s
Board power (TDP) unverified 175 W
Power connectors unverified 1x 8-Pin
Bus interface unverified PCIe 3.0 x16
Display outputs unverified 1x HDMI 2.0b, 3x DisplayPort 1.4a, 1x USB Type-C
DirectX support unverified 12 Ultimate
Vulkan support unverified 1.3
OpenGL support unverified 4.6
Release date unverified 2018-10-17
Release year unverified 2018
Launch MSRP unverified 499 USD