ZLUDA, the famous code porting library known for attempting to break the CUDA moat, has received a major upgrade, as it now supports AMD's ROCm 7.
ZLUDA Acts As a Drop-In Replacement For CUDA, Allowing the Framework to Run on Non-NVIDIA GPUs
NVIDIA's CUDA software ecosystem is built on decades of enabling and developing work, and it is one of the largest moats in the modern-day AI industry. CUDA has evolved into the industry standard for AI frameworks, and with that, various efforts have been made to establish a mechanism that ports CUDA code to non-NVIDIA GPUs. One such tool is ZLUDA. The software has a long history, but for those unaware, it was previously under official development by AMD before becoming an independent project, with Andrzej Janik as the lead developer.
According to a report by Phoronix, ZLUDA has now gained support for AMD's latest software framework, ROCm 7, which will ultimately enable leveraging Team Red's latest software stack after the CUDA code porting. For those unaware, ZLUDA is a drop-in replacement to CUDA code, and the tool essentially intercepts CUDA API calls and redirects them to a different GPU runtime, ultimately allowing for cross-compatibility. If you are wondering why this tool hasn't become mainstream, it's mainly because ZLUDA is still in the development phase. There is no information available on how code porting affects performance.
We do know that firms like Microsoft are attempting to have specific translation layers in place to try to make CUDA universal for their AI workloads, and based on what we have seen in the AI industry, CUDA is the most sought-after toolkit, and probably one of the reasons why NVIDIA has been the dominant compute provider to AI giants. It would be interesting to see whether tools like ZLUDA actually manage to make their way into mainstream AI workloads; however, for now, these are early implementations.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
