Important changes to forums and questions
All forums and questions are now archived. To start a new conversation or read the latest updates go to forums.mbed.com.
10 years, 6 months ago.
floating point resolution/precision
Hi,
Any idea where I can find the epsilon values for double? I need to know the smallest fraction I can use in a double.
Also, is there a version of sin() that takes and rest urns long doubles?
Thanks.
1 Answer
10 years, 6 months ago.
Hi Johan.
Since the floats are IEEE 754 64 bit precision doubles, they have a machine epsilon of b ^-(p-1)/2. Where b is the base and p is the precision (53 for double with an implicit bit). This gives you 1.11e-16. So on average you can count on 16 significant digits.
Also, unless my memory fails me, all C 90 Math routines take and output doubles by default.
Regards,
Alfred