pchardware.org
Categories

NVIDIA · 2016

NVIDIA P40 GPU accelerator

The NVIDIA P40 GPU accelerator is a server-oriented graphics card based on the Pascal architecture, released in 2016.

The NVIDIA P40 GPU accelerator is a server-oriented graphics card based on the Pascal architecture, released in 2016. It features the GP102 chip built on a 16 nm process, with 3840 shading units. The card operates at a base clock of 1303 MHz and a boost clock of 1531 MHz. It is equipped with 24 GB of GDDR5 memory on a 384-bit bus, providing a memory bandwidth of 480 GB/s. The P40 has a TDP of 250 W and requires an 8-pin power connector. It uses a PCIe 3.0 x16 interface and has no display outputs, indicating its use as a compute accelerator in servers. It supports DirectX 12.0, Vulkan 1.2, and OpenGL 4.6. The card was launched on October 7, 2016, with an MSRP of 6999 USD.

Specifications

Architecture unverified Pascal
GPU chip unverified GP102
Process node unverified 16 nm
Shading units unverified 3840
Base clock unverified 1303 MHz
Boost clock unverified 1531 MHz
Memory size unverified 24 GB
Memory type unverified GDDR5
Memory bus unverified 384 bit
Memory bandwidth unverified 480 GB/s
Board power (TDP) unverified 250 W
Power connectors unverified 8-pin
Bus interface unverified PCIe 3.0 x16
Display outputs unverified No display outputs
DirectX support unverified 12.0
Vulkan support unverified 1.2
OpenGL support unverified 4.6
Release date unverified 2016-10-07
Release year unverified 2016
Launch MSRP unverified 6999 USD
Notes unverified No display outputs; intended for server/accelerator use.