Loading...
Discovering amazing AI tools


NVIDIA CUDA 13.1 — a GPU computing toolkit and runtime for accelerating compute and AI workloads, introducing a Tile Programming Model.

NVIDIA CUDA 13.1 — a GPU computing toolkit and runtime for accelerating compute and AI workloads, introducing a Tile Programming Model.
CUDA 13.1 is a release of NVIDIA's GPU computing platform and developer toolkit that accelerates high-performance computing, data-parallel workloads, and AI. It bundles compiler toolchains, runtime libraries, debuggers, samples, and language bindings (including CUDA Python) so developers can write, optimize, and debug GPU kernels. The 13.1 release introduces a Tile Programming Model to let developers work with tiles of data for improved locality and memory access patterns, and includes updated components such as CUDA Python bindings, CUDA-GDB sources for the toolkit, and refreshed sample collections to demonstrate new features and performance techniques. CUDA 13.1 is designed to increase developer productivity while delivering high performance across NVIDIA GPUs.
