What is Nvidia?

What it is

A GPU is a chip built to do many simple calculations at the same time. Drawing graphics needs that, and so does the math behind machine learning. Nvidia supports this second use with CUDA, its software platform for running general computing jobs on its GPUs.

GPUs come as cards that plug into an x86 server or workstation, or as a share of a machine rented from a cloud provider.

What businesses use it for

• AI and machine learning: training models and running them to answer requests.
• 3D and video work: design, rendering and editing software that needs a strong graphics chip.
• Remote desktops: graphics-heavy applications served to people who connect from elsewhere.
• Scientific computing: simulations and analysis that divide into many parallel tasks.

How KawaConnect works with Nvidia

We install and configure Nvidia GPUs in servers and workstations when a workload needs them. That covers fitting the card, checking power and cooling, installing drivers, and making the card available to the software that uses it.

Some are AI workloads like those on our AI and automation page. A GPU can sit in your own rack or be rented by the hour from the cloud. Owning it, on-premises, pays off for steady workloads and for data that must stay in your building. Renting suits occasional jobs.

Learn more

• Nvidia on WikipediaA neutral overview of the company and its products.
• Graphics processing unit on WikipediaA plain explanation of what a GPU is and how it differs from a CPU.
• Nvidia websiteProduct pages and driver downloads for Nvidia graphics cards and data center GPUs.
• Nvidia developer siteDocumentation and tools for people who write software that runs on Nvidia GPUs.