Having implemented floating point in software I can tell you how miserable getting it right (and computing correct ieee status flags) is.
Formally proving correctness must have been miserable.
With the right mindset, there is no misery.
Having implemented floating point in software I can tell you how miserable getting it right (and computing correct ieee status flags) is.
Formally proving correctness must have been miserable.
With the right mindset, there is no misery.