Nvidia introduced the CUDA parallel-computing platform in 2006, helping establish its role in accelerated computing.
CUDA, short for Compute Unified Device Architecture, allows software developers to use Nvidia graphics-processing units for general-purpose computation. Before this approach became widespread, GPUs were mainly associated with rendering images and video games. CUDA made their highly parallel architecture more accessible to scientific, engineering, and machine-learning workloads.
Nvidia announced CUDA in 2006 and released its software development tools alongside compatible graphics hardware. The platform supports programming languages and libraries that let developers divide suitable calculations into many simultaneous operations.
CUDA is not a graphics card and is not a general synonym for artificial intelligence. It is Nvidia's computing platform and software ecosystem. Its importance increased as researchers and companies sought faster computation for neural networks, simulations, data analysis, and other workloads that benefit from massive parallelism.