vosen/ZLUDA
CUDA on non-NVIDIA GPUs observed · 2026-08-28
Health v2 · maintenance only
85/100
- Activity 99
- Release rhythm 58
- Longevity 100
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.
- gap_med: 272.0
- age_days: 2432
- days_rel: 65
- days_push: 7
- n_releases_24m: 3
Adoption not part of the score
14777 stars · 932 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded
ZLUDA is a drop-in replacement layer for CUDA that lets unmodified CUDA applications run on non-NVIDIA GPUs with near-native performance. It implements the CUDA API on top of other GPU runtimes such as AMD's ROCm.
Use cases
- run CUDA applications on AMD GPUs
- run Stable Diffusion on non-NVIDIA hardware
- run llama.cpp with CUDA binaries on AMD GPUs
- avoid rewriting CUDA code for other GPU vendors
- use CUDA-based ML tools without an NVIDIA card
When to choose
- you have CUDA-only software and a non-NVIDIA GPU
- you cannot modify the application source code
- you want near-native performance without porting to ROCm or Vulkan
When to avoid
- you can natively target ROCm, Vulkan, or SYCL instead
- you need guaranteed full CUDA feature coverage or vendor support
- your workload depends on NVIDIA-exclusive features like NVLink or cuDNN edge cases
Facets
library · maturity active
compiler gpu-computing sdk gpu-computing developer-tools machine-learning windows rust cpp cuda-compatibility amd-gpu rocm binary-compatibility gpgpu linux
2 sources
- readme: https://github.com/vosen/ZLUDA · fetched 2026-08-28 · 27cea2ceafd5
- homepage: https://vosen.github.io/ZLUDA/ · fetched 2026-08-29 · a5a68a8a8958
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| vosen/ZLUDA | main | 85 |
For agents
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem