Wouldn’t it be really helpful when you’re trying to optimize part of a script to have a way of timing it? Maybe you want to see which part is really slowing it down. Or, maybe you want to see how much changing the way you calculate something can improve performance.

Well, MATLAB has a function for that. Actually, it’s a pair of functions: tic() and toc().

To use them, you just place tic() before the code to be timed and toc() after it. Then, when you run it the number of seconds elapsed between calling tic() and toc() is output to the command line.

Here’s an example of their use in a code snippet we’ve seen before, the calculation of $e^x$ using a Taylor series expansion:

%% MATLAB snippet to calculate Taylor expansion of the exponential
%% function highlighting use of '``waitbar()``' to generate a progress bar

lim = 5000;   % # of terms in expansion
x = 0.2;   % value of exponent
taylor = 0;

f = waitbar(0,'Progress');   % establish progress bar

tic()
for i = 0:lim
   waitbar(i/lim,f,'Progress');    % update progress bar
   taylor = taylor + x^i/factorial(i);   % add term to expansion
end
toc()

close(f)   % close progress bar

On my computer this takes a full 6.9 seconds. Wow! I wonder what the slow step is. You might think it’s the part where we’re calculating the $i^{th}$ term in the expansion because that’s where all the computational “heavy lifting” is occuring. I’ll comment that line out and see how long it takes:

%  taylor = taylor + x^i/factorial(i);   % add term to expansion

Now it’s 6.1 seconds. Hmmmm. What else could it be? I guess it could be the waitbar() function. Let’s put back the Taylor expansion line and comment out the waitbar() line:

%   waitbar(i/lim,f,'Progress');    % update progress bar

Now it’s only 0.03 seconds!!! I guess there’s a lot going on in updating the waitbar() window every iteration. What if we only update it every 100th iteration?

for i = 0:lim
    if mod(i,100) == 0
        waitbar(i/lim,f,'Progress');    % update progress bar every iteration
    end
   taylor = taylor + x^i/factorial(i);   % add term to Taylor expansion
end

Now it only takes 0.5 seconds. Definitely rendering the waitbar() window is taking the lion’s share of the time.

Hopefully you now see how the simple tic() and toc() functions can be handy in optimizing your scripts!

Read more at the official MathWorks documentation for the tic() and toc() functions.