arlitt at granite.hpl.hp.com
Tue Jan 30 06:50:38 PST 2007
I'll pass along your suggestion to the development team.
On Tue, 30 Jan 2007, Chris Wilson wrote:
> Hi Martin,
> On Mon, 29 Jan 2007, Martin Arlitt wrote:
> > as you know the true median cannot be higher than the maximum. however,
> > httperf trades off a small amount of accuracy in order to reduce the
> > memory required. instead of keeping a list of all the connection times,
> > it keeps a histogram; the median value that is reported is the middle of
> > the bin that contains the median value. for example, if the bin held
> > all observed values from 4 to 5 ms, then the reported median would be
> > 4.5ms, no matter where the observed values fell within the bin.
> Thanks for the explanation. Would it make sense for httperf to report the
> median as "4 ms" or "4.5 +/- 1.0 ms" to avoid giving a false impression of
> the accuracy?
> Cheers, Chris
> Aptivate | http://www.aptivate.org | Phone: +44 1223 760887
> The Humanitarian Centre, Fenner's, Gresham Road, Cambridge CB1 2ES
> Aptivate is a not-for-profit company registered in England and Wales
> with company number 04980791.
> httperf mailing list
> httperf at linux.hpl.hp.com
More information about the httperf