How It Works
CUDA
Nvidia's software platform that lets AI programs run calculations on its graphics chips (GPUs) instead of ordinary processors. Nearly all large AI models are trained on CUDA, which is a big reason Nvidia dominates AI hardware.
Origin
Created at Nvidia; Ian Buck and colleagues published the first public release in 2007, growing out of Buck's Stanford PhD work on the 'Brook' GPU programming language. CUDA originally stood for 'Compute Unified Device Architecture', though Nvidia later dropped the expansion.