This is machine translation

Translated by Microsoft
Mouseover text to see original. Click the button below to return to the English version of the page.

Note: This page has been translated by MathWorks. Click here to see
To view all translated materials including this page, select Country from the country navigator on the bottom of this page.


Apply function to each element of array on GPU


A = arrayfun(FUN, B)
A = arrayfun(FUN,B,C,...)
[A,B,...] = arrayfun(FUN,C,...)


This method of a gpuArray object is very similar in behavior to the MATLAB® function arrayfun, except that the actual evaluation of the function happens on the GPU, not on the CPU. Thus, any required data not already on the GPU is moved to GPU memory, the MATLAB function passed in for evaluation is compiled for the GPU, and then executed on the GPU. All the output arguments return as gpuArray objects, whose data you can retrieve with the gather method.

A = arrayfun(FUN, B) applies the function specified by FUN to each element of the gpuArray B, and returns the results in gpuArray A. A is the same size as B, and A(i,j,...) is equal to FUN(B(i,j,...)). FUN is a function handle to a function that takes one input argument and returns a scalar value. FUN must return values of the same class each time it is called. The input data must be an array of one of the following types: numeric, logical, or gpuArray. The order in which arrayfun computes elements of A is not specified and should not be relied on.

FUN must be a handle to a function that is written in the MATLAB language (i.e., not a MEX-function).

For more detailed information, see Run Element-wise MATLAB Code on GPU. For the subset of the MATLAB language that is currently supported by arrayfun on the GPU, see Supported MATLAB Code.

A = arrayfun(FUN,B,C,...) evaluates FUN using elements of arrays B, C, ... as input arguments with singleton expansion enabled. The resulting gpuArray element A(i,j,...) is equal to FUN(B(i,j,...),C(i,j,...),...). The inputs B, C, ... must all have the same size or be scalar. Any scalar inputs are scalar expanded before being input to the function FUN.

One or more of the inputs B, C, ... must be a gpuArray; any of the others can reside in CPU memory. Each array that is held in CPU memory is converted to a gpuArray before calling the function on the GPU. If you plan to use an array in several different arrayfun calls, it is more efficient to convert that array to a gpuArray before making the series of calls to arrayfun.

[A,B,...] = arrayfun(FUN,C,...), where FUN is a function handle to a function that returns multiple outputs, returns gpuArrays A, B, ..., each corresponding to one of the output arguments of FUN. arrayfun calls FUN each time with as many outputs as there are in the call to arrayfun. FUN can return output arguments having different classes, but the class of each output must be the same each time FUN is called. This means that all elements of A must be the same class; B can be a different class from A, but all elements of B must be of the same class, etc.

Although the MATLAB arrayfun function allows you to specify optional parameter name/value pairs, the gpuArray arrayfun method does not support these options.


  • The first time you call arrayfun to run a particular function on the GPU, there is some overhead time to set up the function for GPU execution. Subsequent calls of arrayfun with the same function can run significantly faster.

  • Nonsingleton dimensions of input arrays must match each other. In other words, the corresponding dimensions of arguments B, C, etc., must be equal to each other, or equal to one. Whenever a dimension of an input array is singleton (equal to 1), arrayfun uses singleton expansion to virtually replicate the array along that dimension to match the largest of the other arrays in that dimension. In the case where a dimension of an input array is singleton and the corresponding dimension in another argument array is zero, arrayfun virtually diminishes the singleton dimension to 0.

    The size of the output array A is such that each dimension is the largest of the input arrays in that dimension for nonzero size, or zero otherwise. Notice in the following code how dimensions of size 1 are scaled up or down to match the size of the corresponding dimension in the other argument:

    R1 = rand(2,5,4,'gpuArray');
    R2 = rand(2,1,4,3,'gpuArray');
    R3 = rand(1,5,4,3,'gpuArray');
    R = arrayfun(@(x,y,z)(x+y.*z),R1,R2,R3);
      2     5     4     3
    R1 = rand(2,2,0,4,'gpuArray');
    R2 = rand(2,1,1,4,'gpuArray');
    R = arrayfun(@plus,R1,R2);
      2     2     0     4
  • Because the operations supported by arrayfun are strictly element-wise, and each element’s computation is performed independently of the others, certain restrictions are imposed:

    • Input and output arrays cannot change shape or size.

    • Functions like rand do not support size specifications. Arrays of random numbers have independent streams for each element.

    For more limitations and details, see Tips and Restrictions.


If you define a MATLAB function as follows:

function [o1,o2] = aGpuFunction(a,b,c)
o1 = a + b;
o2 = o1 .* c + 2;

You can evaluate this on the GPU.

s1 = gpuArray(rand(400));
s2 = gpuArray(rand(400));
s3 = gpuArray(rand(400));
[o1,o2] = arrayfun(@aGpuFunction,s1,s2,s3);
  Name        Size         Bytes  Class

  o1        400x400          108  gpuArray
  o2        400x400          108  gpuArray
  s1        400x400          108  gpuArray
  s2        400x400          108  gpuArray
  s3        400x400          108  gpuArray

Use gather to retrieve the data from the GPU to the MATLAB workspace.

d = gather(o2);

Introduced in R2010b