GPU-Accelerated Libraries

GPU-Accelerated libraries provide highly-optimized algorithms and functions you can incorporate into your applications, with minimal changes to your existing code. Many support drop-in compatibilty to replace industry standard CPU-only libraries such as MKL, IPP, FFTW and widely-used libraries. Some also feature automatic multi-GPU performance scaling.


A simple path to accelerated core solvers, providing up to 10x acceleration in the computationally intense linear solver portion of simulations, and is very well suited for implicit unstructured methods.


NVIDIA cuDNN is a GPU-accelerated library of primitives for deep neural networks, it is designed to be integrated into higher-level machine learning frameworks.


NVIDIA CUDA Fast Fourier Transform Library (cuFFT) provides a simple interface for computing FFTs up to 10x faster, without having to develop your own custom GPU FFT implementation.

IndeX Framework

NVIDIA IndeX Framework is a real-time scalable visualization plug-in for ParaView.


nvGRAPH Analytics Library is a GPU-accelerated graph analytics library.


NVIDIA TensorRT is a high performance neural network inference library for deep learning applications


NVIDIA Performance Primitives is a GPU accelerated library with a very large collection of 1000's of image processing primitives and signal processing primitives.


FFmpeg is a popular open-source multi-media framework with a library of plugins that can be applied to various parts of the audio and video processing pipelines.


GPU-accelerated CHOLMOD is part of the SuiteSparse linear algebra package by Prof. Tim Davis. SuiteSparse is used extensively throughout industry and academia.

CULA Tools

GPU-accelerated linear algebra library by EM Photonics, that utilizes CUDA to dramatically improve the computation speed of sophisticated mathematics.


A collection of next gen linear algebra routines. Designed for heterogeneous GPU-based architectures. Supports current LAPACK and BLAS standards.

IMSL Fortran Numerical Library

Developed by RogueWave, a comprehensive set of mathematical and statistical functions that offloads work to GPUs.


A collection of dense and sparse direct solvers which deliver significant acceleration for Computer Vision, CFD, Computational Chemistry, and Linear Optimization applications


NVIDIA CUDA Sparse (cuSPARSE) Matrix library provides a collection of basic linear algebra subroutines used for sparse matrices that delivers over 8x performance boost.


NVIDIA CUDA BLAS Library (cuBLAS) is a GPU-accelerated version of the complete standard BLAS library that delivers 6x to 17x faster performance than the latest MKL BLAS.


Comprehensive, open source GPU function library. Includes functions for math, signal and image processing, statistics, and many more. Interfaces for C, C++, Java, R and Fortran.


The CUDA Random Number Generation library performs high quality GPU-accelerated random number generation (RNG) over 8x faster than typical CPU only code.

CUDA Math Library

An industry proven, highly accurate collection of standard mathematical functions, providing high performance on NVIDIA GPUs.


A powerful, open source library of parallel algorithms and data structures. Perform GPU-accelerated sort, scan, transform, and reductions with just a few lines of code.


A GPU-accelerated C++ framework for High-Throughput Sequence Analysis for both short and long read alignment.


Gunrock is a GPU library for graph analytics that couples an expressive, powerful programming model with a high-performance implementation.


Accelerate video compression with the NVIDIA Video Codec SDK. This SDK includes documentation and code samples that illustrate how to use NVIDIA’s NVENC and NVDEC hardware in GPUs to accelerate encode, decode, and transcode of H.264 and HEVC video formats.


HiPLAR (High Performance Linear Algebra in R) delivers high performance linear algebra (LA) routines for the R platform for statistical computing using the latest software libraries for heterogeneous architectures.


OpenCV is the leading open source library for computer vision, image processing and machine learning, and now features GPU acceleration for real-time operation.

Geometry Performance Primitives(GPP)

GPP is a computational geometry engine that is optimized for GPU acceleration, and can be used in advanced Graphical Information Systems (GIS), Electronic Design Automation (EDA), computer vision, and motion planning solutions.


A library for sparse iterative methods with special focus on multi-core and accelerator technology such as GPUs.

Triton Ocean SDK

Triton provides real-time visual simulation of the ocean and bodies of water for games, simulation, and training applications.