10 raised to the power

When I multiply two matrices element wise (which is probably not important to my question), some numbers appear as 3.400000000000001e+03 instead of simply 3400. Why?

Answer 1

[Star Strider]

Put simply: format if you are viewing them in the Command Window. This is likely also due to floating-point approximation error, as described in Floating-Point Numbers.

Otherwise, change the format descriptor in fprintf formatSpec or similar function calls to change the way they are displayed in those function calls.

There is not enough detail in the Question to provide an exact response.

.

[Walter Roberson]

format long g
delta = 0.1;
C = 0; for K = 1 : 34000; C = C + delta; end
C

C = 
            3399.999999998

C - 3400

ans = 
     -1.99679561774246e-09

1/10 is not exactly representable in binary floating point. When you compute with non-integer floating point numbers, unless you restrict yourself to pure powers of 2 such as 0.25, then you should expect that the result may be different than you would expect mathematically.

[Star Strider]

@Walter Roberson — Thank you!

.