Field Guide · organization

Also known as: NVIDIA Corporation

NVIDIA is an American company founded in 1993 that designs GPUs and the CUDA software platform for general-purpose GPU computing.1

1993 founded 1999 GeForce coins "GPU" 2007 CUDA GPU compute now AI / HPC accelerators
NVIDIA moved from graphics to general compute: the 1999 GeForce popularized the term GPU, CUDA in 2007 turned that GPU into a programmable parallel processor, and the same hardware now leads the AI and HPC accelerator market.

Overview

NVIDIA was co-founded and is led by Jensen Huang. It popularised the term “GPU” with its GeForce line and later turned the GPU into a general parallel processor by releasing CUDA in 2007. That move made NVIDIA hardware the standard platform for deep learning, and the company is now a leading supplier of AI and high-performance computing accelerators.2

Beyond add-in cards, NVIDIA designs Tegra systems-on-chip and the Jetson line of embedded AI modules, bringing GPU-accelerated computing to robotics, cameras, and edge devices.

Product lines

NVIDIA’s silicon now spans from gaming cards to data-center accelerators and edge modules:

Product line What it is
GeForce Consumer GPUs for gaming and graphics
Data-center GPUs Accelerators for AI training and HPC
Jetson Embedded GPU modules for edge AI
CUDA The software platform that unlocks GPU compute

Where it fits

GPUs excel at the massively parallel arithmetic that DSP and machine learning both demand. An NVIDIA card can accelerate heavy SDR signal processing or run on-device classification, and a Jetson AI accelerator can do GPU-accelerated work at the edge — for example, near an antenna where sending raw IQ back to a server would be impractical. Like most leading-edge designs, NVIDIA’s chips are fabricated by a foundry such as TSMC.

Sources

  1. Nvidia — Wikipedia, for the company’s history and significance. 

  2. NVIDIA — the company’s official site, for products and CUDA. 

See also