# JuliaGPU/CUDA.jl

CUDA programming in Julia.

Repository: https://github.com/JuliaGPU/CUDA.jl
Canonical: https://ross.abutalabs.com/products/cudajl
Homepage: https://juliagpu.org/backends/cuda/
Language: Julia
License: NOASSERTION
License Family: other
Topics: julia, cuda, gpu, hacktoberfest
Last push: 2026-08-26T19:10:15+00:00

## Health v2 (maintenance only)
Score: 99/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 100
- inputs: {"age_days": 2553, "days_push": 7, "days_rel": 12, "gap_med": 10.0, "n_releases_24m": 35}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1424, forks 279 (observed 2026-08-28T04:04:41.341391+00:00)

## What it is
CUDA.jl is the main Julia package for programming NVIDIA CUDA GPUs, offering a high-level CuArray abstraction, a compiler for writing CUDA kernels in Julia, and wrappers for CUDA libraries. It has been developed since 2014 and aims for CUDA C-level performance without requiring a separate CUDA toolkit installation.

## Use cases
- run array computations on an NVIDIA GPU from Julia
- write custom CUDA kernels in Julia instead of C
- accelerate scientific computing workloads with GPUs
- offload matrix operations to the GPU without GPU programming experience
- interface directly with CUDA APIs and libraries from Julia

## When to choose
- you program in Julia and want GPU acceleration on NVIDIA hardware
- you want a high-level array abstraction with automatic GPU offloading
- you need CUDA C-level performance while writing kernels in Julia
- you want automatic CUDA toolkit installation without manual setup

## When to avoid
- you use AMD or Intel GPUs rather than NVIDIA
- you need GPU programming in Python, C++, or another language
- you have no NVIDIA driver available on your system

## Facets
- artifact type: library
- maturity: stable
- function: gpu-computing, compiler, sdk, benchmarking
- domain: gpu-computing, machine-learning, data-science, programming-languages
- platform: windows
- tags: cuda, nvidia, julia, gpu-programming, kernels, cuaray, high-performance-computing, linux, macos, gpu, docker

## Member repositories
- JuliaGPU/CUDA.jl (main) score 99

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:41.341391+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:37:33.462317+00:00, confidence not recorded.
  - readme: https://github.com/JuliaGPU/CUDA.jl (fetched 2026-08-28T04:04:41.341391+00:00, sha e60d5215358f)
  - homepage: https://juliagpu.org/backends/cuda/ (fetched 2026-08-29T11:49:30.395172+00:00, sha 2f89761929c4)
  - site_page: https://cuda.juliagpu.org/stable (fetched 2026-08-29T11:49:30.398442+00:00, sha a326789f64a1)
- Data as of 2026-08-30T08:39:29.467469+00:00.