Overview
Examples
Screenshots
Comparisons
Applications
Download
Documentation
Tutorials
Bazaar
Status & Roadmap
FAQ
Authors & License
Forums
Funding Ultimate++
Search on this site











SourceForge.net Logo

ArrayFire

ArrayFire package includes simple ArrayFire benchmarks and demos.

 

ArrayFire is a library for fast GPU computing, supporting both Nvidia CUDA and OpenCL devices, and it is open source (BSD 3-clause).

 

It includes matrix algebra, algorithms, image processing, etc. classes and functions.

 

Benchmarks and demos are:

Pi        Pi number benchmark

Matrix    NxN matrix product benchmark

Vectorize    Different strategies to do operations between vectors

Demo    Basic matrix algebra demos

 

 

You can download ArrayFire sources from GitHub or the binaries from the ArrayFire page.

 

In the last case you will have to register (no cost, no personal data required) and after installing you will get a "include" and "lib" folder, that also includes the required dll.

 

Just set:

the "include" folder in your Setup/Build methods/Include directories

the "lib" folder in your Setup/Build methods/Lib directories

the "lib" folder in your Setup/Build methods/Path - executable directories

 

and in the "Main package configuration" set AF_CPU (ArrayFire libraries use your CPU cores), AF_OPEN_CL  (ArrayFire libraries use your GPU through OpenCL) or AF_CUDA (ArrayFire libraries use your GPU through Nvidia CUDA).

 

Finally install the GPU drivers:

If you have a Nvidia graphic card, you can install CUDA

For the rest of graphic cards as AMD/ATI or Intel, you can install OpenCL from their vendor download web

 

Last edit by koldo on 01/06/2016. Do you want to contribute?. T++