Re: unicode performance penalty

Andreas Piesk wrote:
> applicant2006_at_rambler.ru schrieb:
>
> > I'm in the process of migrating our database to Unicode and have
> > discovered that it works 4 to 5 times faster than non-unicode database.
>
> unicode is 4-5 times faster? you mean slower, right?

Right. a spelling error -:).

>
> > Oracle docs says that Unicode db should perform same or better than a
> > non-unicode one. Perhaps someone can verify my configuration and point
> > possible problem areas ?
> >
> > Server : Oracle 9.2.0.1.0
>
> i found 9.2.0.4 the first usable 9i version. upgrade possible?

Yes, sure, but I'd like to get a feeling first what other people's experiences are.

>
> > Client : java thin jdbc ( ojdbc14.jar )
>
> have you tried 10g jdbc driver? 9i jdbc has several problems. we
> generally use the latest 10g jdbc oci drivers, even with 9i databases.
>

I trield both classes12.jar and odbc14.jar with the same result. I didn't get an impression that it's a client's problem.

> > Before the conversion to Unicode we could perform about 6000 inserts
> > per second.
> > After the conversion we can perform only about 1800-2000 inserts per
> > second, even if I triple RAM used by Oracle.
>
> more RAM wouldn't help to speed up the inserts. have you determined
> where all the time was spent?

It doesn't use RAM that much. The bottleneck is definetely CPU. Oracle uses 100% of CPU time. I have to admit that it sounds logical - someone has to deal with all this extra Unicode data, but on the other side java already sends Unicode, so no conversion is necessary... Uhhh, I don't know....

>
> regards,
> -ap
Received on Tue Jan 09 2007 - 01:18:33 CST