How to fast process a calculation within a matrix without using a loop?

I have a huge matrix A, and what i need to do is to go through each of the matrix element and do this below simple calculation:

[m n] = size(A);
for i=1:m
    for j=1:n
        B(i,j) = 10^-A(i,j);
    end
end

The thing is that it is slow. Is there a way I can avoid the loop and make it faster?

Thank you!

[per isakson]

Pre-allocation!

B = nan(size(A));

[Leon]

Thank you. What does this do?

It is likely my data might contain NaNs.

[Rik]

Whenever the size of an array increases, Matlab has to create a copy in memory. This takes a lot of time if you repeat that very often. Your original code would cause the output to be extended m+n-1 times.

Pre-allocation would solve this issue by creating an array in advance of the correct size. If array processing is not possible, it is generally worth it to pre-allocate the output array.

[Leon]

Got you!

Many thanks!

Answer 1

[Rik]

Yes, use the power of array processing in Matlab:

B=10.^-A;

[Leon]

Many thanks!

Will that be a problem, if my data also contains NaNs?

[Rik]

It will do the same as your loop for NaN values, so the B matrix will contain NaN everywhere where A has a NaN.