Another option is simply to put "tic" on top of your code and "toc" at the end.

Use the cputime function.

At beginning of your code: tstart = cputime;

At the end: disp(['Execution completed in: ' cputime-tstart]);

Output is in seconds.

Another option is simply to put "tic" on top of your code and "toc" at the end.

Robert ,thank you very much for your help.

ok, thanks Ulf Mallast

Mathworks recommends that the 'tic / toc' solution suggested by #Ulf Mallast should be preferred to 'elapsed time':

>> help etime

etime Elapsed time.

etime(T1,T0) returns the time in seconds that has elapsed between

vectors T1 and T0. The two vectors must be six elements long, in

the format returned by CLOCK:

T = [Year Month Day Hour Minute Second]

Time differences over many orders of magnitude are computed accurately.

The result can be thousands of seconds if T1 and T0 differ in their

first five components, or small fractions of seconds if the first five

components are equal.

Note: When timing the duration of an event, use the TIC and TOC

functions instead of CLOCK or etime. These latter two functions are

based on the system time which can be adjusted periodically by the

operating system and thus might not be reliable in time comparison

operations.

Example:

This example shows two ways to calculate how long a particular FFT

operation takes. Using TIC and TOC is preferred, as it can be

more reliable for timing the duration of an event:

x = rand(800000, 1);

t1 = tic; fft(x); toc(t1) % Recommended

Elapsed time is 0.097665 seconds.

t = clock; fft(x); etime(clock, t)

ans =

0.1250

See also tic, toc, clock, cputime, datenum.

Reference page in Help browser

doc etime

thank you Fernando

Indeed the commands "tic" and "toc" are the simplest ones. However, if you want a more detailed report, analysing the time and the execution percentage for each line of your code, the best command is "profile". It creates a report in HTML format that can be opened automatically at your standard browser.

To access the browser automatically, Matlab will need some toolbox files, like:

\toolbox\matlab\general\ibrowse.dll

\toolbox\matlab\general\web.m

For additional information about Matlab Profiling, check

http://www.mathworks.com/help/matlab/matlab_prog/profiling-for-improving-performance.html

ok thank you Wederley .

Kouba, I have used both functions: tic/toc and cputime. If you also have the same code implemented in C/C++ and compare the execution time of both codes using the tic/toc functions, you will be able to see a big difference. This happens because of the difference in the details of implementation. The use of the tic/toc functions is more indicated when you are running only your MATLAB code. They work based on the processor clock and yield a more precise outcome.