On Fri, Jan 20, 2006 at 02:04:58AM +0300, Mike Belopuhov wrote:
> On Thu, Jan 19, 2006 at 15:17 -0700, Tobias Weingartner wrote:
> > > some test results are available at
> > > http://crypt.org.ru/~mkb/pmap/
> >
> > These look interesting. Can you describe better what each axis is,
> > and what the tests you are running are?
>
> markus@ have gave me some benchmarking program in private mail...
> i don't even know if i can publish it (;
>
> Y axis is a number of time elapsed, and X axis displays a number of
> processes, is suppose (it's all markus with his progz!

only that particular benchmark shows a win.
average use does not produce any noticable improvement
because most of the gain is lost on tree rebalancing
where it not needed.

perhaps a better tree is the answer.

cu
--
paranoic mickey (my employers have changed but, the name has remained)