CUDA

NVIDIA's parallel computing platform and API for GPU acceleration.

CUDA enables developers to leverage NVIDIA GPUs for tasks like AI inference, as defaulted in vLLM's v0.20.0. It powers high-throughput applications in frameworks competing with ROCm. Its dominance influences hardware-agnostic projects to support alternatives for broader accessibility.


Category: framework · Also: Compute Unified Device Architecture · Mentioned in 3 Cortex outputs