PRODUCT

CUDA

CUDA is a parallel computing platform and API developed by NVIDIA. It lets software use the GPU for general-purpose computation and is the standard runtime for modern AI training.

Maker: NVIDIA
Introduced: 2007

PRODUCTCUDA

Trails Articles

Key facts

What we know

Maker: NVIDIA
Introduced: 2007

Forest Trails

Connections in the forest

Made by
NVIDIA(Company)
Runs on
GPU(Technology)
Enables
Machine Learning(Concept)GPU Computing(Concept)Artificial Intelligence(Technology)

Referencing articles

5 articles

Companies6 minUpdated Jun 10, 2025

Why NVIDIA matters

It is not just the silicon. CUDA and the software ecosystem around it are what keep AI workloads on NVIDIA.

Technology6 minUpdated Apr 30, 2025

GPU computing, explained

The GPU is not a faster CPU. It is a differently shaped processor, optimised for problems that a CPU is bad at.

Technology6 minUpdated Mar 15, 2025

What is CUDA?

CUDA lets ordinary C-like code target thousands of GPU cores in parallel. That is the whole trick.

Companies8 minUpdated Feb 25, 2025

The story of NVIDIA

NVIDIA started by selling graphics cards to gamers. Three decades later, its chips train almost every serious AI model.

Technology8 minUpdated Jan 30, 2025

How ChatGPT works

ChatGPT is a large language model plus a chat interface plus a training procedure that teaches the model to be helpful.

Continue exploring

Neighboring entities in the forest.

Amazon(Company)Amazon Web Services(Product)AMD(Company)Andrej Karpathy(Person)Anthropic(Company)Apple(Company)Apple Silicon(Product)Artemis Program(Concept)

Or search the whole forest →