"Howard J. Rogers" wrote:
>
> Anecdotally, and not strictly benchmarked as such, but when I've replaced
> Windows with Linux (ie, on the same machine), I've had tests come out 17-20%
> faster. I was gobsmacked at the time, and rather worried that all my
> previous testing had been measuring Windows and not Oracle!
>
> I've refused to install Oracle on Windows ever since. Unless I really,
> really have to ;-)
>
> Regards
> HJR
My own [informal] experience supports your observation as well with idle
X services on the machine (eg: on KDE login screen) and roughly 15% more
when running with the X server off (eg: typically run level 3).
/Hans
Received on Fri Nov 14 2003 - 18:03:42 CST
