We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.

Saved items Sign in

NVIDIA

NVIDIA Tesla P40 24GB GDDR5 graphics card

SKU CH-P6DC-3RAY4

A 24GB GDDR5 accelerator with 3840 CUDA cores delivering 12 TFlops single-precision compute for deep learning and scientific workloads

  • Memory TypeGDDR5
  • 3840 CUDA cores accelerate parallel workloads
  • 24 GB GDDR5 memory handles large datasets
  • 346 GB/s bandwidth feeds compute pipelines
  • 12 TFlops peak single-precision performance speeds simulations
  • PCIe x16 interface fits HPE ProLiant servers

NVIDIA Tesla P40 24GB GDDR5 Accelerator

The NVIDIA Tesla P40 is a compute accelerator designed for data centres and HPC clusters that need high single-precision throughput. It targets deep learning inference, scientific simulation and professional visualisation workloads. The board delivers 12 TFlops peak performance from 3840 CUDA cores. It ships as a new pull with a power cord included.

PCIe x16 integration for HPE ProLiant servers

The card occupies a single PCIe x16 slot and draws auxiliary power through its onboard connectors. It is validated for select HPE ProLiant server families, allowing drop-in deployment without chassis modification. The dual-slot form factor leaves adjacent slots free for networking or storage controllers. No external power brick or proprietary cabling is required.

24GB GDDR5 memory with 346 GB/s bandwidth

The 24GB GDDR5 frame buffer provides ample capacity for large neural networks, high-resolution textures and scientific datasets. With ECC disabled the memory subsystem sustains 346 GB/s, keeping all 3840 cores fed during dense matrix operations. Workloads that exceed 24GB will spill to system memory, reducing throughput.

Highlights

  • 3840 CUDA cores accelerate parallel workloads
  • 24 GB GDDR5 memory handles large datasets
  • 346 GB/s bandwidth feeds compute pipelines
  • 12 TFlops peak single-precision performance speeds simulations
  • PCIe x16 interface fits HPE ProLiant servers

Specifications

Memory TypeGDDR5

Questions about this item

Does this card fit a standard PCIe x16 slot?

Yes, it uses a PCIe x16 interface for installation in compatible servers.

What power connection does the card require?

A power cord is included in the box for the necessary auxiliary power.

How much memory bandwidth is available for compute workloads?

The board delivers 346 GB/s of memory bandwidth with ECC disabled.

Also in this aisle

People compared these

Same shelf, same checkout — eight coins and a 30-minute rate lock.