The Compute Unified Device Architecture provides programming models, compilers, libraries, developer tools, and optimized kernels for Nvidia graphics processors. Its mature ecosystem supports artificial intelligence, scientific computing, graphics, and other parallel workloads.
The platform’s advantage comes from accumulated software and expertise as well as hardware. Alternative chips must offer usable compilers, libraries, kernels, debugging, and deployment paths or find ways to generate that support efficiently.
Acronyms and aliases
CUDA acronymNvidia CUDA variant
Related terms
Frequently asked questions
What does CUDA stand for?
CUDA stands for Compute Unified Device Architecture, Nvidia’s platform for general parallel computing on its graphics processors.
Why is CUDA described as a moat?
Its long-developed libraries, tools, optimized code, compatibility, and engineering expertise make switching to another accelerator costly.
