NVIDIA Labs today released CUDA-Oxide 0.1, an experimental project implementing a Rust-to-CUDA compiler. This new initiative is designed to improve the capabilities of using the Rust programming language for developing CUDA kernels for NVIDIA GPUs. The release integrates Rust's memory safety and performance benefits into the CUDA development ecosystem. The project, which Phoronix described as "greatly improving the capabilities of using the Rust programming language for developing CUDA kernels for NVIDIA GPUs," is in its initial experimental phase. It provides developers with a new avenue for writing high-performance GPU code while leveraging modern language features from Rust.