News for hackers. Stuff that matters.
Re:Intel Underestimates Error Bounds by 1.3 Quintillion by Danieru on Thursday October 09, 2014 @07:16PM (#8435490)
> Storing those 4 values takes ~29 bits.

Just storing those 4 values in 4 bits is more efficient.

Checking that a bit is set is also a simple AND + CMP. Which beats out a DIV + CMP.

Sorry I just had not considered this isn't common knowledge outside C or C++ land.

Re:Intel Underestimates Error Bounds by 1.3 Quintillion by Thrymr on Thursday October 09, 2014 @07:54PM (#8435584)
> Just storing those 4 values in 4 bits is more efficient.

That doesn't really work for version numbers > 1. They're not just flags.

Re:Intel Underestimates Error Bounds by 1.3 Quintillion by njr123 on Thursday October 09, 2014 @08:12PM (#8435631)
How do you store a version number in a single bit?
Re:Intel Underestimates Error Bounds by 1.3 Quintillion by rjaco31 on Friday October 10, 2014 @03:31AM (#8436640)
Well the bit n represents the version n, if this bit is at 1, this version is supported, otherwise it's not.
Re:Intel Underestimates Error Bounds by 1.3 Quintillion by ajuc on Friday October 10, 2014 @04:05AM (#8436709)
A few bits per value (depending how big you predict they will get). One bit per value means you can only increment version of each function once :)