iro-cuda-ffi
| Crates.io | iro-cuda-ffi |
| lib.rs | iro-cuda-ffi |
| version | 0.2.1 |
| created_at | 2026-01-18 13:59:14.537917+00 |
| updated_at | 2026-01-18 15:24:52.757533+00 |
| description | IRO CUDA FFI - A minimal, rigid ABI boundary for Rust to orchestrate nvcc-compiled CUDA kernels |
| homepage | https://github.com/iro/iro-cuda-ffi |
| repository | https://github.com/iro/iro-cuda-ffi |
| max_upload_size | |
| id | 2052386 |
| size | 265,229 |
Tribe IRO (tribe-iro)
documentation
https://docs.rs/iro-cuda-ffi
README
iro-cuda-ffi
Core IRO CUDA FFI crate.
This crate provides the ABI types, error handling, streams, events, and device memory management for Rust code orchestrating nvcc-compiled kernels.
CUDA 12.0 or later is required.
See the workspace README.md for the full overview, kernel examples, and
benchmark guidance.