tracel-ai/cubecl

*One codebase. More hardware choices.** Write your compute kernels in Rust. CubeCL translates them for different GPUs and CPUs, so you can reuse the same logic across hardware backends. CubeCL is a Rust language extension, a Just-in-Time compiler, and a set of runtimes for writing high-performance compute kernels. A single #[cube] Rust function is compiled on demand and runs on CUDA, HIP, Metal, Vulkan, WebGPU, or CPU, while still using the best instructions each platform offers. The programming model is low-level by design, so that a single kernel can reach peak performance on every backend i…