NVlabs

NVlabs/cuda-oxide

cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign language bindings, just Rust.

3.0K stars 227 forks 3.0K watchers Rust Apache License 2.0
async compiler-backend cuda gpu heterogeneous-computing high-performance-computing nvidia programming-languages rust
View on GitHub Website
1 Open Issue Need Help Last updated: Jul 31, 2026

Open Issues Need Help

View All on GitHub
Value-threaded WGMMA accumulator lowering for pipelined use (wait_group<N>, multiple groups in flight) about 2 hours ago
enhancement help wanted perf IR-lowering cuda-feature
NVlabs/cuda-oxide
3.0K

cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign language bindings, just Rust.

Rust
#async#compiler-backend#cuda#gpu#heterogeneous-computing#high-performance-computing#nvidia#programming-languages#rust