Top Welcoming Open-Source Cuda Repositories

mKernel: fast multi-node, multi-GPU fused kernels

Graphics Processing Units Molecular Dynamics

GPU accelerated decision optimization

A plugin to use Nvidia GPU in PySCF package

cuGraph - RAPIDS Graph Analytics Library

cuVS - a library for vector search and clustering on the GPU

A high-performance, GPU-accelerated library for key computational chemistry tasks, such as molecular similarity, conformer generation, and geometry relaxation.

GPU Accelerated t-SNE for CUDA with Python bindings

A GPU-Accelerated First-Order LP Solver

RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.

An efficient and research-friendly Gaussian Splatting framework described in the CVPR'26 paper "Faster-GS: Analyzing and Improving Gaussian Splatting Optimization"

GPU-accelerated triangle mesh processing

Tile primitives for speedy kernels

CUDA Library Samples

how to optimize some algorithm in cuda.

[MICRO'23, MLSys'22] TorchSparse: Efficient Training and Inference Framework for Sparse Convolution on GPUs.

RWKV-7 7.2B fp16 15000+ tps decoding @ single 5090

NCCL Tests

LLM training in simple, raw C/CUDA

This package contains the original 2012 AlexNet code.

Deformable ConvNets V2 (DCNv2) in PyTorch

[CVPR 2023 Highlight] Neural Kernel Surface Reconstruction

Instant neural graphics primitives: lightning fast NeRF and more

Mirage Persistent Kernel: Compiling LLMs into a MegaKernel

Code from the "CUDA Crash Course" YouTube series by CoffeeBeforeArch

DeepGEMM: clean and efficient BLAS kernel library on GPU