Get the latest tech news

CubeCL: GPU Kernels in Rust for CUDA, ROCm, and WGPU


Multi-platform high-performance compute language extension for Rust. - tracel-ai/cubecl

With CubeCL, you can program your GPU using Rust, taking advantage of zero-cost abstractions to develop maintainable, flexible, and efficient compute kernels. We currently have an highly optimized matrix multiplication module, leveraging Tensor Cores on NVIDIA hardware where available, while gracefully falling back to basic instructions on other platforms. However, since the algorithmic behavior may depend on the vectorization factor, CubeCL allows you to access it directly in the kernel when needed, without any performance loss, using the comptime system!

Get the Android app

Or read this on Hacker News

Read more on:

Photo of Rust

Rust

Photo of rocm

rocm

Photo of CUDA

CUDA

Related news:

News photo

Sapphire: Rust based package manager for macOS

News photo

AMD ROCm 6.4 Adds SPIR-V Linking Support To HIP

News photo

Pretty State Machine Patterns in Rust (2016)