Ecosystem of libraries and tools for writing and executing fast GPU code fully in Rust.

Up to 100x faster strings for C, C++, CUDA, Python, Rust, Swift, JS, & Go, leveraging NEON, AVX2, AVX-512, SVE, GPGPU, & SWAR to accelerate search, hashing, sorting, edit distances, sketches, and memory ops 🦖

Safe rust wrapper around CUDA toolkit

Rusty wrapper for the CUDA Driver API

Demo project showing a single Rust codebase running on CPU and directly on GPUs

Pure Rust + CUDA LLM inference engine

Doing advent of code with CUDA and rust.

Ruby gem for running state-of-the-art language models locally. Access LLMs, embeddings, rerankers, and NER models directly from Ruby using Rust-powered Candle with Metal/CUDA acceleration.

Cuq: A MIR-to-Coq Framework Targeting PTX for Formal Semantics and Verified Translation of Rust GPU Kernels

Multilingual neural TTS (6 languages: JA/EN/ZH/ES/FR/PT, code supports SV) — C++, C#, Rust, Go, Python, npm (WASM). VITS + Prosody, streaming, CUDA/CoreML/DirectML. pip install piper-plus | npm install piper-plus | cargo install piper-plus-cli

Rust binding to CUDA APIs

Thrust, CUB, TBB, AVX2, AVX-512, CUDA, OpenCL, OpenMP, Metal, and Rust - all it takes to sum a lot of numbers fast!

LLaMa 7b with CUDA acceleration implemented in rust. Minimal GPU memory needed!

Clean, simple, hackable ml framework in rust. Supports metal and cuda

A learning-focused, high-performance tensor computation library built from scratch in Rust, featuring automatic differentiation and CPU/CUDA backends.

Safe CUDA cuDNN wrapper for the Rust language.

rust-cuda working group

HowTo: Compile CUDA with nvcc, and link to Rust

MD5 hash cracking with CUDA and Rust, implemented from scratch

Asynchronous CUDA for Rust.

a high-performance computing solution designed to accelerate matrix operations using Nvidia's CUDA technology with Rust's ndarray data structure.

A Rust wrapper for Darknet, an open source neural network framework written in C and CUDA.

Safe CUDA cuBLAS wrapper for the Rust language.

Native and Private ML inference engine, embeddings, classification, reranking, search, and text generation. Rust core with C# bindings. No Python, no ONNX, no CUDA.

Experiments with CUDA and Rust

Ultra high-performance secp256k1 ECC engine | Python, Node.js, Rust, Go, C#, Swift, Java bindings | CUDA, Metal, OpenCL GPU | ECDSA, Schnorr, FROST, MuSig2, BIP-352 | 15+ platforms

A fancy GPU path-tracer for photorealistic lighting and rendering written in C++, Rust, and CUDA C

Pure-Rust inference engine for Qwen3-ASR speech recognition models (0.6B & 1.7B) using candle with Metal/CUDA acceleration

Let's jump into CUDA development with Rust

PyTorch-equivalent ML framework in pure Rust — 22 crates, CUDA GPU, biometrics, IR detection, LLMs, ONNX, distributed training