I look forward to the day we can start using quad-precision numbers, but was disappointed to see that in the specification, only 15 bits out of a whopping 128 were assigned to the exponent as shown by Wikipedia. That's only 4 bits more than we had for doubles!

This allows for great precision, but we sacrifice the high number range. In my years of programming, I have always found extremely high numbers to be pretty useful for various calculations and purposes, and sometimes the 'double' doesn't cut the mustard. It also helps simplify many algorithms when we don't need to worry about number overflow.

I was just wondering if the seemingly heavy bias towards precision had a good reason. Are there practical purposes for such extreme precision?

  • $\begingroup$ Have you tried reading the relevant standard? $\endgroup$ Apr 29, 2015 at 14:11
  • $\begingroup$ @YuvalFilmus: Not sure quite what you mean. Is there an official document which goes into more detail about why the spec is what it is? $\endgroup$
    – Dan W
    Apr 29, 2015 at 14:43
  • $\begingroup$ Exactly. The document defining this format might have some relevant information. $\endgroup$ Apr 29, 2015 at 14:57
  • $\begingroup$ @YuvalFilmus: Yes I had a look for it, but couldn't find anything. $\endgroup$
    – Dan W
    Apr 29, 2015 at 15:02


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.