[cairo] Cairo 1.3 performance loss

[Behdad Esfahbod]

On Tue, 2007-02-06 at 11:10 +0200, Jorn Baayen wrote:
> 
> > These directly lead to a simple algorithm: while total time spent <
> X
> > seconds or total number of runs < Y runs, do repeat the test.
> 
> This sounds good, but how would you normalize the recorded numbers?
> Dividing by number of runs will in many cases throw us back to
> sub-centisecond precision.

Why exactly?  If you measure something a million times and divide the
total time by a million, you have drastically reduced the error caused
by your timer's limited precision.

What you have not reduced is the overhead of running the test, but that
was not the point here.

-- 
behdad
http://behdad.org/

"Those who would give up Essential Liberty to purchase a little
 Temporary Safety, deserve neither Liberty nor Safety."
        -- Benjamin Franklin, 1759