GoKawiilDeveloper Bridges ZLUDA and ROCm to Run CUDA Workloads on AMD GPUs in Windows
An independent developer known as Speedstu built a PowerShell-based toolkit that combines ZLUDA with AMD's HIP/ROCm SDK, allowing CUDA-only applications to run on AMD GPUs in Windows without dual-booting or virtual machines. The setup successfully mapped CUDA's driver API and libraries like cuBLAS, cuSPARSE, and cuFFT to AMD equivalents, and was tested by training a small reinforcement-learning model on a Radeon RX 9060 XT using unmodified CUDA code.
GoKawiil's interpretation of the reporting above, not reported fact.
Even as AMD expands native ROCm support on Windows, many AI tools and research repositories remain hardcoded to CUDA, locking out non-NVIDIA hardware. This project shows a practical, if unofficial, path for AMD GPU owners to access that ecosystem without needing Linux or virtualization, potentially broadening the practical utility of AMD's consumer GPUs for AI experimentation.
- The toolkit merges ZLUDA with AMD's HIP/ROCm SDK to translate CUDA calls on Windows.
- Core CUDA libraries like cuBLAS, cuSPARSE, and cuFFT work, but cuDNN support is not yet functional.
- Currently only the Radeon RX 9060 XT is officially supported, with a working proof-of-concept using a small neural network.
ASUS ROG Strix G16 Gaming Laptop — The ASUS ROG Strix G16 Gaming Laptop is a powerhouse designed for gamers and creators alike, now capable of handling CUDA-exclusive workloads thanks to recent developments in AMD GPU support. Its robust hardware and compatibility make it an excellent choice for those looking to run demanding AI and CUDA-based applications on AMD hardware without sacrificing performance or flexibility.
See ASUS ROG Strix G16 Gaming Laptop on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: tomshardware.com — Zak Killian, 2026-09-14
Published there as: “Solo dev enables running CUDA on AMD hardware in Windows, getting multiple CUDA libraries running on a gaming Radeon RX 9060 XT GPU in Windows — CUDA-exclusive workloads on AMD hardware in Windows possible without virtualization or dual-booting”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.