John Kasunich wrote: > > I respectfully disagree on this part. Counts are counts, and > integers can be accumulated indefinitely with no error. Floating > point math can be imprecise, and accumulating "counts times scale", > where scale is a float, could result in accumulating errors. > > <snip> > If you are accumulating counts as an integer, and then > multiplying by the scale factor, the error will always be > 1 part in 2^24, regardless of how far you travel. > Yes, if what we are talking about is the rolling over of a smaller counter (possibly in hardware) to a wider position value, then a raw count value has to be maintained as an integer, and converted to float every cycle. Different floating point formats will have different thresholds of minimum guaranteed precision, but accumulating a series of products will be guaranteed to eventually accumulate significant error.
Jon ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
