VUDA is a header-only library based on Vulkan that provides a CUDA Runtime API interface for writing GPU-accelerated applications. It conforms to the specification of the CUDA runtime and is based on the Vulkan API. The VUDA wiki provides detailed information on setup, compilation, deviations from CUDA, and implementation details. All VUDA functionality can be accessed by including vuda.hpp and using the vuda:: namespace, or by utilizing vuda_runtime.hpp which wraps and redirects all CUDA functionality. The provided usage example demonstrates how to allocate memory on the device, copy arrays to the device, run a Vulkan shader module or a CUDA kernel, and copy the resulting data back to the host. Finally, the example illustrates freeing memory on the device.
