Open-source gpu-computing software

386 open-source libraries, tools and applications tagged gpu-computing, ranked by how closely gpu-computing is their primary purpose, then by adoption and maintenance health. Each entry links to a page with sources, use cases and when to choose or avoid it.

386 products, primary matches first, then adoption-weighted; health v2 shown.

Product	Health v2	Stars	Maturity
Taichi Taichi Lang is an open-source domain-specific language embedded in Python for high-performance parallel numerical computation, using a JIT …	60	28390	active
hashcat Hashcat is a high‑performance password recovery utility that runs on CPUs, GPUs and distributed systems. It supports hundreds of hash algor…	79	26999	active
Dao-AILab/flash-attention Official implementation of FlashAttention, FlashAttention-2, -3, and -4: fast and memory-efficient exact attention kernels for GPUs. It acc…	92	25116	active
mlc-ai/mlc-llm MLC LLM is a machine learning compiler and high-performance deployment engine for large language models, built on TVM-based ML compilation.…	61	23176	active
NVIDIA/open-gpu-kernel-modules The source release of NVIDIA's open Linux GPU kernel modules, released per driver version and built with make. It must be paired with match…	93	17411	active
tracel-ai/burn Burn is a Rust-based tensor library and deep learning framework supporting training and inference through a unified API. It JIT-compiles te…	86	16008	active
gpujs/gpu.js GPU.js is a JavaScript acceleration library for GPGPU that automatically transpiles JavaScript functions into shader language (GLSL) and co…	98	15505	active
vosen/ZLUDA ZLUDA is a drop-in replacement layer for CUDA that lets unmodified CUDA applications run on non-NVIDIA GPUs with near-native performance. I…	83	15103	active
NVIDIA/TensorRT-LLM NVIDIA's TensorRT-LLM is a Python library and runtime for defining and serving Large Language Models with state-of-the-art inference optimi…	83	14851	active
Eclipse Deeplearning4j Eclipse Deeplearning4J is an open-source deep learning framework and ecosystem for the JVM, including the ND4J linear algebra library, the …	76	14269	active
NVIDIA/TensorRT NVIDIA TensorRT is an SDK for high-performance deep learning inference on NVIDIA GPUs, comprising an inference compiler that converts train…	93	13362	stable
deepseek-ai/FlashMLA FlashMLA is DeepSeek's library of optimized CUDA attention kernels implementing Multi-head Latent Attention (MLA), including dense and toke…	64	13063	active
cupy/cupy CuPy is a NumPy/SciPy-compatible array library for GPU-accelerated computing with Python, running on NVIDIA CUDA or AMD ROCm. It acts as a …	92	12362	stable
Taskflow Taskflow is a header-only C++ library for building task-parallel programs using task dependency graphs, with a work-stealing executor for C…	84	12188	stable
Numba Numba is an open-source NumPy-aware JIT compiler that translates a subset of Python and NumPy code into fast machine code using LLVM. It su…	94	11175	stable
Syllo/nvtop NVTOP is an htop-like ncurses task monitor for GPUs and AI accelerators, supporting NVIDIA, AMD, Intel, Apple, Huawei Ascend, Qualcomm Adre…	85	11034	active
cumulo-autumn/StreamDiffusion StreamDiffusion is a Python pipeline for real-time interactive diffusion-based image generation, achieving 100+ fps on modern GPUs. It opti…	18	10817	active
NVIDIA/cutlass CUTLASS is NVIDIA's collection of CUDA C++ template abstractions and Python DSLs for implementing high-performance GEMM and related linear …	98	10580	active
Tiiny-AI/PowerInfer PowerInfer is a CPU/GPU hybrid LLM inference engine that exploits activation locality to run large language models fast on consumer-grade h…	60	9814	active
NVIDIA/cudf cuDF is a GPU-accelerated DataFrame library for tabular data processing, part of NVIDIA's RAPIDS suite. It provides a pandas-compatible Pyt…	95	9780	stable
OneFlow OneFlow is an open-source deep learning framework written in C++ with a PyTorch-like Python API, focused on scalable and efficient distribu…	44	9437	active
NVIDIA/apex NVIDIA-maintained PyTorch extension providing utilities for easy mixed precision and distributed training. It offers up-to-date CUDA and C+…	75	9000	active
deepseek-ai/DeepGEMM DeepGEMM is a high-performance CUDA BLAS kernel library for NVIDIA tensor cores, providing FP8, FP4, and BF16 GEMMs plus fused MoE and othe…	82	8937	active
tile-ai/tilelang TileLang is a Pythonic domain-specific language and compiler (built on TVM) for writing high-performance GPU/CPU/accelerator kernels such a…	90	8658	active
zeux/meshoptimizer A C/C++ library providing algorithms to optimize triangle meshes for GPU rendering, including vertex cache optimization, overdraw reduction…	94	8518	stable
Dynamo NVIDIA Dynamo is an open-source, datacenter-scale distributed inference serving framework that orchestrates inference engines like vLLM, SG…	88	8345	active
kroma-network/tachyon Tachyon is a modular zero-knowledge (ZK) proving backend written in C++17, designed to let developers implement arbitrary proving schemes w…	20	7640	active
XuehaiPan/nvitop nvitop is an interactive NVIDIA GPU process viewer and monitoring tool built on NVML, offering a richer alternative to nvidia-smi with an h…	89	7184	active
flashinfer-ai/flashinfer FlashInfer is a GPU kernel library and kernel generator for LLM inference, providing unified APIs for attention, GEMM, and MoE operations w…	92	6728	active
linkedin/Liger-Kernel Liger Kernel is a collection of Triton kernels designed to make LLM training faster and more memory-efficient, offering drop-in replacement…	91	6664	active
PufferLib PufferLib is a fast, open-source reinforcement learning library that trains tiny, super-human models in seconds, achieving 1M+ environment …	87	6476	active
GPUStack GPUStack is an open-source GPU cluster manager for AI model serving that orchestrates inference engines like vLLM, SGLang, and TensorRT-LLM…	88	5844	active
Newton Newton is an open-source, GPU-accelerated physics simulation engine built on NVIDIA Warp, targeting roboticists and simulation researchers.…	87	5813	active
Slang Slang is a shading language and compiler for building large, modular, high-performance GPU shader codebases. It compiles to D3D12, Vulkan, …	96	5752	active
NVIDIA/cuml NVIDIA cuML is a GPU-accelerated machine learning library offering scikit-learn-style estimators that run on NVIDIA GPUs via CUDA. It also …	91	5292	active
ProjectPhysX/FluidX3D FluidX3D is a high-performance lattice Boltzmann method computational fluid dynamics (CFD) application written in C++/OpenCL. It runs on al…	90	5290	active
NCCL NVIDIA's Collective Communication Library (NCCL) is a C++ library providing topology-aware, high-bandwidth inter-GPU communication primitiv…	99	5165	stable
ArrayFire ArrayFire is a general-purpose tensor/numerical computing library for C, C++, and Python that accelerates array operations on GPUs (CUDA, O…	65	4903	stable
llm-d llm-d is a Kubernetes-native distributed LLM inference serving stack that orchestrates model servers like vLLM and SGLang across clusters. …	83	4891	active
spaceandtimefdn/blitzar Blitzar is a C++ library (with Rust bindings) that accelerates zero-knowledge proof algorithms on CPU and GPU using CUDA. It was built by S…	84	4863	active
HAMi HAMi (Heterogeneous AI Computing Virtualization Middleware) is a CNCF Incubating, Kubernetes-native GPU virtualization and scheduling middl…	91	4691	active
openxla/xla XLA (Accelerated Linear Algebra) is an open-source machine learning compiler that optimizes models from frameworks like PyTorch, TensorFlow…	77	4605	active
st-tech/ppf-contact-solver A GPU-accelerated contact solver for physics-based simulations of shells (cloth), solids, rods, rigid bodies, and sand, originally develope…	87	4513	active
HIP HIP is a C++ runtime API and kernel programming language for AMD GPUs that mirrors the NVIDIA CUDA programming interface. It enables develo…	90	4406	active
gpustat gpustat is a simple command-line utility for querying and monitoring NVIDIA GPU status, offering a compact alternative to nvidia-smi. It di…	66	4391	active
danveloper/flash-moe An application for running large language models locally on small laptops, written in Objective-C. It enables on-device inference of big mo…	44	4238	active
zml/zml ZML is a production LLM inference stack written in Zig, built on MLIR and OpenXLA, that compiles models to run at peak performance across N…	69	4136	active
uxlfoundation/oneDNN oneDNN is an open-source cross-platform performance library providing optimized building blocks (primitives) for deep learning applications…	99	4067	stable
AnswerDotAI/gpu.cpp A lightweight, header-only C++20 library for portable low-level GPU compute built on Dawn's WebGPU implementation. It provides a minimal AP…	74	3990	active
thu-ml/SageAttention SageAttention is a family of quantized attention kernels (INT8/FP8/FP4) that accelerate transformer inference 2-5x over FlashAttention with…	38	3956	active
NVIDIA/k8s-device-plugin NVIDIA's official Kubernetes device plugin that exposes NVIDIA GPUs as schedulable resources in Kubernetes clusters. It supports GPU sharin…	94	3901	stable
thu-ml/TurboDiffusion TurboDiffusion is a Python framework that accelerates end-to-end video diffusion model generation by 100-200x using SageAttention, Sparse-L…	58	3851	active
HazyResearch/ThunderKittens ThunderKittens is a C++/CUDA framework of tile-based primitives for writing fast deep learning GPU kernels. It embeds natively into CUDA so…	68	3754	active
mit-han-lab/llm-awq AWQ is a Python library implementing Activation-aware Weight Quantization for compressing large language models to low-bit (INT3/4) weights…	41	3638	active
NVIDIA/cuda-python NVIDIA's official Python package for accessing the CUDA platform, providing Pythonic APIs and low-level bindings to CUDA C APIs, parallel a…	95	3399	active
software-mansion/TypeGPU TypeGPU is a modular TypeScript toolkit for WebGPU that provides advanced type inference and lets developers write WGSL shaders directly in…	93	3267	active
onnx/onnx-tensorrt A C++ parser library and backend that converts ONNX models into TensorRT engines for high-performance GPU inference. It is maintained by NV…	91	3238	active
Sergio0694/ComputeSharp ComputeSharp is a .NET library that lets developers write compute and pixel shaders in C# and run them in parallel on the GPU via DirectX 1…	73	3173	stable
BinomialLLC/basis_universal Basis Universal is an open-source LDR/HDR GPU supercompressed texture codec and transcoding system supporting .KTX2 and .basis file formats…	93	3120	active
pytorch/TensorRT Torch-TensorRT is a compiler library that accelerates PyTorch model inference on NVIDIA GPUs using TensorRT. It supports just-in-time compi…	93	3009	active
vLLM vllm-ascend is a community-maintained hardware plugin that enables vLLM to run large language model inference on Huawei Ascend NPUs. It imp…	83	2998	active
turboderp/exllama ExLlama is a standalone Python/C++/CUDA implementation of the Llama architecture optimized for running 4-bit GPTQ quantized weights on mode…	30	2947	active
Luce-Org/lucebox Lucebox is an open-source LLM inference engine written in C++ with custom CUDA/HIP kernels, focused on speculative decoding and speculative…	60	2941	active
TheHPXProject/hpx HPX is a C++ Standard Library for parallelism and concurrency that implements C++ standard parallel facilities and extends them to distribu…	68	2936	stable
elixir-nx/nx Nx is an Elixir library providing multi-dimensional arrays (tensors) and numerical definitions with pluggable backends. The monorepo also i…	98	2915	active
rednote-machine-learning/RedKnot RedKnot is a long-context LLM inference acceleration library built on SGLang, using head-classified KV reuse, offline KV storage with RoPE …	56	2840	active
diku-dk/futhark Futhark is a purely functional, statically typed data-parallel array programming language in the ML family, developed at the University of …	95	2809	active
pytorch/xla PyTorch/XLA is a Python package that connects the PyTorch deep learning framework to XLA devices such as Google Cloud TPUs via the XLA deep…	63	2801	active
xdit-project/xDiT xDiT is a scalable inference engine for Diffusion Transformers (DiTs) that enables parallel deployment across multiple GPUs and machines. I…	91	2732	active
lightseekorg/tokenspeed TokenSpeed is a high-performance LLM inference engine designed for agentic workloads, aiming for TensorRT-LLM-level performance with vLLM-l…	77	2732	active
CVCUDA/CV-CUDA CV-CUDA is an open-source GPU-accelerated library of computer vision and image processing operators built on CUDA, with C++ and Python APIs…	92	2729	active
qualcomm/aimet AIMET (AI Model Efficiency Toolkit) is a Python library from Qualcomm providing advanced quantization and compression techniques for traine…	99	2723	active
Kokkos Kokkos Core is a C++ programming model providing abstractions for parallel execution and memory management to write performance-portable ap…	98	2707	stable
open-mpi/ompi Open MPI is an open-source, standards-conformant implementation of the Message Passing Interface (MPI) for high-performance computing, deve…	77	2663	stable
openlake-project/openlake OpenLake is a high-performance distributed storage engine written in Rust (built on io_uring, RDMA, and GPUDirect) designed to feed GPUs du…	78	2662	active
NVlabs/LongLive LongLive is an NVIDIA research framework providing parallel training and inference infrastructure for real-time long video generation, usin…	60	2639	active
meta-pytorch/torchrec TorchRec is a PyTorch domain library for building recommendation systems at scale. It provides distributed sharding of large embedding tabl…	95	2615	active
KomputeProject/kompute Kompute is a general-purpose GPU compute framework built on Vulkan that works across vendor GPUs (AMD, NVIDIA, Qualcomm, etc.) with both C+…	63	2577	active
mirage-project/mirage Mirage Persistent Kernel (MPK) is a compiler and runtime that transforms multi-GPU LLM inference into a single fused megakernel, reducing i…	81	2560	active
CUDA Core Compute Libraries CUDA Core Compute Libraries (CCCL) is NVIDIA's unified repository for Thrust, CUB, and libcudacxx, providing high-performance C++ abstracti…	99	2541	active
lupinemachines/lupine LUPINE is a GPU-over-IP bridge that lets machines without local GPUs attach NVIDIA GPUs running on remote servers over the network. It expo…	91	2452	active
tracel-ai/cubecl CubeCL is a Rust language extension, JIT compiler, and runtime set for writing high-performance GPU compute kernels in plain Rust. A single…	83	2437	active
Vortex GPGPU Vortex is a full-stack open-source RISC-V GPGPU supporting OpenCL, with configurable cores, warps, and threads, plus graphics and tensor co…	78	2286	active
666DZY666/micronet micronet is a Python library for deep neural network model compression and deployment built on PyTorch. It provides quantization (QAT, PTQ,…	38	2265	active
MFEM MFEM is a lightweight, modular C++ library for finite element discretization of PDEs, supporting arbitrary high-order element spaces, adapt…	85	2258	stable
PKU-YuanGroup/Helios Helios is a 14B autoregressive diffusion model for real-time, minute-scale video generation supporting text-to-video, image-to-video, and v…	56	2158	active
NVIDIA/cutile-python cuTile Python is a tile-based programming language and library from NVIDIA for writing parallel kernels that run on NVIDIA GPUs, compiling …	62	2151	active
cvxpy/cvxpylayers CVXPYlayers is a Python library for constructing differentiable convex optimization layers in PyTorch, JAX, and MLX using CVXPY. It solves …	90	2141	active
GPUOpen-Effects/FidelityFX-FSR2 AMD FidelityFX Super Resolution 2 (FSR 2) is an open-source, high-quality temporal upscaling solution that reconstructs high-resolution fra…	23	2080	stable
inducer/pycuda PyCUDA is a Python wrapper giving Pythonic access to Nvidia's CUDA parallel computation API, including GPUArray multidimensional arrays and…	78	2052	active
chapel-lang/chapel Chapel is a modern open-source programming language designed for productive parallel computing at scale, with first-class support for task …	90	2047	active
NUS-HPC-AI-Lab/VideoSys VideoSys is an open-source Python library providing easy and efficient infrastructure for video generation, supporting training, inference,…	31	2025	active
Strix Halo AI Toolboxes A collection of pre-built Toolbx/Docker containers ('toolboxes') for running llama.cpp LLM inference on AMD Ryzen AI Max 'Strix Halo' integ…	64	1998	active
NVIDIA/Stable-Diffusion-WebUI-TensorRT An NVIDIA extension for the Stable Diffusion Web UI (Automatic1111) that accelerates image generation using TensorRT-optimized engines on R…	19	1985	active
develsoftware/GMinerRelease GMiner is a closed-source GPU cryptocurrency miner for NVIDIA and AMD cards supporting algorithms such as Ethash, ProgPoW, KAWPOW, Equihash…	64	1975	active
siliconflow/onediff OneDiff is an out-of-the-box acceleration library for diffusion models, providing PyTorch compilation tools and optimized GPU kernels. It i…	44	1965	active
AdaptiveCpp/AdaptiveCpp AdaptiveCpp (formerly hipSYCL/Open SYCL) is an independent, community-driven C++ compiler platform for heterogeneous programming models inc…	72	1951	active
deepseek-ai/TileKernels TileKernels is a Python library of optimized GPU kernels for LLM operations, written in the TileLang DSL. It provides kernels for MoE routi…	58	1949	active
flexflow/flexflow-train FlexFlow Train is a deep learning framework that accelerates distributed DNN training by automatically searching for efficient parallelizat…	67	1903	active
NVIDIA Data Center GPU Manager (DCGM) NVIDIA's GPU metrics exporter for Prometheus, built on top of DCGM (Data Center GPU Manager). It exposes GPU telemetry such as clock speeds…	96	1879	active