conversion of double to int

i had an array of 1*10 double named Q(i). I want to find modulo function of these 10 values with 256. But am getting wrong answers. May be its the problem of types. Will you please help me to solve?

[Walter Roberson]

Could you give some examples of the values and what you are observing ?

[SNEHA P S]

   Q(7) = 4.078e+09

And i would like to mod it using equation

s(i) = mod(Q(i),256)

and i am getting

 s(7) = 25.8600

But when i am trying to do the same in the command window using the exact numerical value ie., 4.078e+09, am getting it right. And its get wrong when am using Q(7) instead of value,

[Adam]

And what is the problem? That seems like the right answer to me.

[SNEHA P S]

sorry, Q(7) value is like 4.078e+09. And when use mod function with 256 the exact value is 128. And am geeting 25.8600.

[Walter Roberson]

Give the command

format long g

And then display the Q value again

Answer 1

[Walter Roberson]

2^53 = 9007199254740992 is the largest integer expressed in double precision that you can reliably take mod() 10 of. After that, the difference between adjacent representable double precision numbers becomes more than 1.

You can extend the range by using uint64(), allowing you to get to 18446744073709551615 . Beyond that you would need to use the Symbolic Toolbox, or use one of the extended precision File Exchange contributions.

[SNEHA P S]

when iam trying to do the same in the command window using the exact numerical value ie., 4.078e+09, am getting it right. And its get wrong when am using Q(7) instead of value.

[Jan]

@SNEHA P S: Then compare the two values:

Q(7) - 4.078e+09

Note that the short format for the command window does not let you distinguish 4078000000 and 4078000001 or 4077999999.