What could be a possible reason of getting a different result from .mex file(which is converted from .m file)?

I wrote a code with matlab language that involves a lot of iterations. To enhance the speed of the code, I converted the m function file to mex file and solution building process was successful. The problem is that I get the different result from mex file. What could be the possible reason of this situation? (I use version 2014a)

Answer 1

[Walter Roberson]

Round-off errors. Even a very slightly different way of writing an expression can give notably different results.

http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F