Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.

RDNA-native LLM inference engine in Rust.

CubeK: high-performance multi-platform kernels in CubeCL

Rust system bindings for AMD ROCm HIP runtime used by CubeCL

软件、补丁、NixOS 模块与 AI 编码助手的技能合集。

Specialised ML library — YaneuraOu(Shogi AI) focused fork of bullet-shogi

Local AMD GPU LLM inference engine (HIP/ROCm) — quantized GGUF, Q4_0/K-quants, flash attention, KV cache. ~515 tok/s decode on Qwen2.5-0.5B Q4_0 (RX 7900 XT).

Silicon Fingerprinting & Hardware OSINT Engine leveraging GPU-accelerated telemetry analysis and microscopic clock-skew variance. Built with Rust, C, and HIP/ROCm.

Heterogeneous Compute Cascade (HCC) — distributed 400B-parameter MoE inference across dual AMD Ryzen AI MAX+ 395 'Strix Halo' workstations via USB4.

Rust-first AMD GPU simulation and sensor framework built on ROCm-Oxide.

Rust-first GPU SDK for AMD ROCm: Rust-authored kernels, HSACO generation, typed bindings, HIPRTC, Vulkan demos, and diagnostics.

Generate vanity OpenPGP keys with a chosen key ID prefix, at billions of hashes per second on AMD and NVIDIA GPUs

tensor- and kernel-acceleration layer built on top of Lamina