ArrayFire is a high performance software library for parallel computing with an easy-to-use API. ArrayFire abstracts away much of the details of programming parallel architectures by providing a high-level container object, the Array, that represents data stored on a CPU, GPU, FPGA, or other type of accelerator. This abstraction permits developers to write massively parallel applications in a high-level language where they need not be concerned about low-level optimizations that are frequently required to achieve high throughput on most parallel architectures.

Supported data types

ArrayFire provides one generic container object, the Array on which functions and mathematical operations are performed. The Array can represent one of many different basic data types:

  • F32 real single-precision (float)
  • C32 complex single-precision (cfloat)
  • F64 real double-precision (double)
  • C64 complex double-precision (cdouble)
  • B8 8-bit boolean values (bool)
  • S32 32-bit signed integer (int)
  • U32 32-bit unsigned integer (unsigned)
  • U8 8-bit unsigned values (unsigned char)
  • S64 64-bit signed integer (intl)
  • U64 64-bit unsigned integer (uintl)
  • S16 16-bit signed integer (short)
  • U16 16-bit unsigned integer (unsigned short)
  • F16 16-bit floating point number (half::f16)

Most of these data types are supported on all modern GPUs; however, some older devices may lack support for double precision arrays. In this case, a runtime error will be generated when the array is constructed.

If not specified, Arrays are created as single precision floating point numbers (F32).

Creating and populating an Array

ArrayFire Array's represent memory stored on the device. As such, creation and population of an array will consume memory on the device which cannot freed until the array object goes out of scope. As device memory allocation can be expensive, ArrayFire also includes a memory manager which will re-use device memory whenever possible.

Arrays can be created using one of the array constructors. Below we show how to create 1D, 2D, and 3D arrays with uninitialized values:

let garbageVals = Array::new_empty(Dim4::new(&[3, 1, 1, 1]), DType::F32);

However, uninitialized memory is likely not useful in your application. ArrayFire provides several convenient functions for creating arrays that contain pre-populated values including constants, uniform random numbers, uniform normally distributed numbers, and the identity matrix:

// Create an array filled with constant value of 2.0 of type floating point
// The type of Array is infered from the type of the constant argument
let cnst = constant(2.0f32, Dim4::new(&[5, 5, 1, 1]));
println!("Create a 5-by-3 matrix of random floats on the GPU");
let dims = Dim4::new(&[5, 3, 1, 1]);
let a = randu::<f32>(dims);

As stated above, the default data type for arrays is F32(32-bit floating point number) unless specified otherwise.

ArrayFire Arrays may also be populated from data found on the host. For example:

let values: [u32; 3] = [1u32, 2, 3];
let indices = Array::new(&values, Dim4::new(&[3, 1, 1, 1]));

Properties of an Array

ArrayFire provides several functions to determine various aspects of arrays. This includes functions to print the contents, query dimensions, and determine various other aspects of arrays.

The print function can be used to print arrays that have already been generated or any expression involving arrays:

let values: [f32; 3] = [1.0, 2.0, 3.0];
let indices = Array::new(&values, Dim4::new(&[3, 1, 1, 1]));

The dimensions of an array may be determined using either a Dim4 object or by accessing the dimensions directly using the Dim4::get and Dim4::numdims functions:

let values: [f32; 3] = [1.0, 2.0, 3.0];
let dims: Dim4 = Dim4::new(&[3, 1, 1, 1]);
let indices = Array::new(&values, dims);
println!("Dims {:?} with dimensions {}", dims.get(), dims.ndims());

In addition to dimensions, arrays also carry several properties including methods to determine the underlying type and size (in bytes). You can even determine whether the array is empty, real/complex, a row/column, or a scalar or a vector. For further information on these capabilities, we suggest you consult the full documentation on the Array.

Writing math expressions using ArrayFire

ArrayFire features an intelligent Just-In-Time (JIT) compilation engine that converts expressions using arrays into the smallest number of CUDA/OpenCL kernels. For most operations on Arrays, ArrayFire functions like a vector library. That means that an element-wise operation, like c[i] = a[i] + b[i] in C, would be written more concisely without indexing, like c = a + b. When there are multiple expressions involving arrays, ArrayFire's JIT engine will merge them together. his "kernel fusion" technology not only decreases the number of kernel calls, but, more importantly, avoids extraneous global memory operations.

Our JIT functionality extends across C API boundary and only ends when a non-JIT function is encountered or a synchronization operation is explicitly called by the code.

ArrayFire provides hundreds of functions for element-wise operations. All of the standard operators (e.g. +,-,*,/) are supported as are most transcendental functions (sin, cos, log, sqrt, etc.). Here are a few examples:

let num_rows: u64 = 5;
let num_cols: u64 = 3;
let dims = Dim4::new(&[num_rows, num_cols, 1, 1]);
let a = randu::<f32>(dims);
let b = randu::<f32>(dims);
let c = a + b;

//Example of *Assign traits
let mut d = randu::<f32>(dims);
let e     = constant(1f32, dims);
d += e;


Like all functions in ArrayFire, indexing is also executed in parallel on the OpenCL/CUDA device. To index Arrays you may use one or a combination of the following functions:

Please see the indexing page for several examples of how to use these functions.

Access to Array memory on the host

Memory in af::Arrays may be accessed using the Array::host() method. The host function copies the data from the device and makes it available in a standard slice or similar container on the host. As such, it is up to the developer to manage any memory returned by host.

Bitwise operators

In addition to supporting standard mathematical functions, Arrays that contain integer data types also support bitwise operators including and, or, and shift etc. Operator traits for Array as well as separate functions are also defined to support various use cases.

let dims = Dim4::new(&[5, 3, 1, 1]);
let a = randu::<bool>(dims);
let b = randu::<bool>(dims);


let c = &a | &b; //Borrowing to avoid move of a and b, a | b is also valid
let d = bitand(&a, &b, false);


Where to go for help?