Skip to content

Latest commit

 

History

History
55 lines (35 loc) · 4.91 KB

README.md

File metadata and controls

55 lines (35 loc) · 4.91 KB

awesome-gpgpu

A curated list of awesome GPGPU (CUDA/OpenCL/Vulkan) resources

CUDA

Libraries

C++

  • cuda-api-wrappers - Thin, unified, C++-flavored wrappers for the CUDA APIs [github eyalroz/cuda-api-wrappers]
  • Thrust - The C++ parallel algorithms library [github NVIDIA/thrust]

Tutorials

OpenCL

Libraries

C++

  • Arrayfire - A General purpose GPU library [github arrayfire/arrayfire]
  • Boost.Compute - A C++ GPU Computing Library for OpenCL [github boostorg/compute]
  • Chlorine - Dead Simple OpenCL [github Polytonic/Chlorine].
  • clpp - Lightweight and easy to use the OpenCL C++ wrapper in single file [github matszpk/clpp].
  • EasyCL - Easy to run kernels using OpenCL [github hughperkins/EasyCL]
  • EasyOpenCL - The easiest way to get started with OpenCL! [github Gladdy/EasyOpenCL]
  • VexCL - VexCL is a C++ vector expression template library for OpenCL/CUDA [github ddemidov/vexcl]

Rust

  • Emu - Emu is a high-level language that compiles to OpenCL and provides a procedural macro for embedding code in Rust [github calebwin/emu]

Applications

  • Bullet - Real-Time Physics Simulation (Bullet 3 includes the optional work-in-progress GPU pipeline.) [github bulletphysics/bullet3]
  • CLBlast - Modern, lightweight, performant and tunable OpenCL BLAS library written in C++11 [github CNugteren/CLBlast]
  • DeepCL - OpenCL library to train deep convolutional neural networks [github hughperkins/DeepCL]
  • libclsph - OpenCL based GPU accelerated SPH fluid simulation library [github libclsph/libclsph]
  • MACE - Deep learning inference framework optimized for mobile heterogeneous computing platforms [github XiaoMi/mace]
  • OpenCL-caffe - OpenCL version of caffe developed by AMD research lab [github amd/OpenCL-caffe]
  • Sol-R - CUDA/OpenCL-based realtime ray-tracer [github favreau/Sol-R]

Tutorials

Vulkan

Libraries

C++

  • Kompute - The general purpose GPU compute framework for cross vendor graphics cards [github KomputeProject/kompute]