How to do accurate float calculations?

[duckwithanokhat]

In my C program I'm doing simple float subtract and addition. The problem is if I do something like 25.00 - 24.99 my float variable would get something like 0.0099.... . While that is sufficient for displaying, I need the number stored to be 0.01. Is there a way to do more accurate float calculations?

[2FA]

%f2

The two indicates that you want it to the hundredths.

[duckwithanokhat]

1 minute ago, DeadEyePsycho said:

%f2

The two indicates that you want it to the hundredths.

Doesn't that only work with displaying?

[Guest]

10 minutes ago, duckwithanokhat said:

Doesn't that only work with displaying?

trunc (a*100/100)

[2FA]

2 minutes ago, deXxterlab97 said:

trunc (a*100/100)

trunc simply truncates the number instead of actually rounding it.

[Guest]

2 minutes ago, DeadEyePsycho said:

trunc simply truncates the number instead of actually rounding it.

Then replace trunc with round. 

[2FA]

#include <math.h>
  
float example = 0.32423523;

  
float rounded = roundf(example*100)/100  
  
#The constant 100 can be adjust for other places, ex being 10 = tenths
#roundf will round to the nearest number

 

[fizzlesticks]

Use a double. Or a long double. 

Or if you actually want real accuracy, convert everything to integers to do the calculations and convert back when outputting the results.

You could also use an arbitrary precision library that can do accurate calculations to whatever precision you tell it to. I don't know any C ones off the top of my head but they shouldn't be hard to find.

[DevBlox]

More on this topic if you're passing by and are curious why this happens. There is a nice video for it.