ArrayFire is a High Performance software library for parallel computing with an easy-to-use API. Its array based function set makes parallel programming simple.
ArrayFire's multiple backends (CUDA, OpenCL and native CPU) make it platform independent and highly portable.
A few lines of code in ArrayFire can replace dozens of lines of parallel computing code, saving you valuable time and lowering development costs.
This repository contains the test data used for testing ArrayFire
The files are in the following format:
number of inputs
4-dimensional size of input 0
...
4-dimensional size of input n
number of results
total elements of results 0
...
total elements of results n
test data for input 0
...
test data for input n
test data for output 0
...
test data for output n
Contributions of any kind are welcome! Please refer to this document to learn more about how you can get involved with ArrayFire.
email: technical@arrayfire.com