The ZLUDA toolkit has been released, allowing CUDA applications to run on AMD GPUs.
The ZLUDA project has prepared an open implementation of CUDA technology for AMD GPUs, allowing unmodified CUDA applications to run with performance close to that of applications run without layers. The released toolkit provides binary compatibility with existing CUDA applications compiled using the CUDA compiler for NVIDIA GPUs. The implementation works on top of AMD's developing ROCm stack and the HIP (Heterogeneous-computing Interface for Portability) runtime. Code […]
