I did a benchmark test a while ago , comapring
WS , OL, SE . Using hp9000 , 1 GB ram , 2 CPU's .
The test consisted of a simulation running a 4gl order entry
application using benchmarking tool called Mercury Loadrunner.
I run 10, 20 , 30, 40 and 50 user tests , bottom line was
the difference was minimal OL was 1.01868 % faster then SE and
same as WS . Note this is a OLTP app , DSS may be different ...
Question how many CPU's do you have , my theory is that
on 2 CPU machine there is little benefit , things change as you get into
large databases and many CPU's .... I know PDQ is great for DSS ...
Mariusz
MARTIN <ramp@arrakis.es> wrote in article <35088D80.32CE70DB@arrakis.es>...
> After the tuning of on Line Work Group Server 7.20UC2 in a Siemens
> machine with 256 i have compared performance : S.E and OWGS have the
> same performance .
>
> I can not understad . Do you?
>
> I have all in on line : buffer o.k , kaio , access to two disk , small
> check point , temp dbspaces , raw devices
>