AMD Introduces a New GPU Technology – ZLUDA

AMD is taking a significant leap forward in GPU software development by introducing a solution that combines the ROCm technology with NVIDIA’s CUDA binary files.

This new solution, known as ZLUDA, allows existing CUDA applications to run directly on AMD Radeon graphics cards without the need for modifications to the original source code.

Some key information about this new technology includes:

ZLUDA: The project responsible for compatibility is named ZLUDA. It was initially developed to bring CUDA support to Intel integrated graphics cards but has now been adapted to work with Radeon cards through AMD’s HIP/ROCm stack.

Binary Compatibility: Thanks to ZLUDA, many existing CUDA workloads can run smoothly on AMD graphics cards without involving programmers. It acts as a library replacement, making it easier for end-users to run CUDA-supported software.

Performance Tests: Performance tests for ZLUDA implementation have shown promising results. CUDA software and renderers worked seamlessly with ZLUDA, and rendering in Blender 4.0 optimized for CUDA now runs faster on AMD Radeon graphics cards than on the native ROCm/HIP port, reducing rendering time by 10-20% depending on the scene.

Open Source: The ZLUDA project is dual-licensed under Apache 2.0 and MIT licenses. Rust language enthusiasts will be pleased to know that it is used in the implementation for Radeon cards.

GitHub Repository: If you’re curious and want to explore the new ZLUDA open-source code for Radeon cards, you can find it on the GitHub platform.

This new technology potentially expands AMD’s reach in the GPU market and promotes healthy competition. It’s exciting to see advancements that benefit both developers and end-users.

Frequently Asked Questions

The source of the article is from the blog tvbzorg.com