Re: tuning an simple insert-statement

"andrewst" <member14183_at_dbforums.com> wrote in message news:3338580.1062860886_at_dbforums.com...
>
> Originally posted by Pascal Ziegler
>
> > my task is to compare the performance of a function in cache with a
>
> > similar implementation in oracle. the insert-statements in the
> > function
>
> > are dynamic (not hard-coded), but for my measurement I used a
> > hard-coded
>
> > statement. cache needs about 70 microseconds to insert the values...
>
> > oracle is about 3-4 times slower. i know that this measurement
> > can't be
>
> > very accurate and there are a lot of other things that influence my
>
> > measurement. but the point is, (after severeal measurements)
> > that oracle
>
> > is always slower. i am a oracle-rookie, i have to do this
> > comparison for
>
> > my degree disertation (university of zurich). now i'm looking for
>
> > possibilities to improve the performance of oracle. i'm not
> > looking for
>
> > the final solution... but just points where i can start to tune
> > my oracle.
>
> >
>
> > what else:
>
> > - there are no triggers firing.
>
> > - i havent traced my function... but I thought that my statement
> > is not
>
> > that complex that i can find the problem with tracing... (?)
>
> >
>
> > by the way: is this a right presumption: I can't improve the
> > performance
>
> > of an insert-statement, when i build a data-cartridge (with an own
>
> > index-structure). i think with an index it even gets slower...?
>
> >
>
> > thank you for the help! :-)
>
> >
>
> > - pascal
>
> >
>
> >
>
> > Ron Reidy wrote:
>
> > > What is slow? How long does it take? Are you performing 100s
> > of 1000s
>
> > > of these hard-coded things? Are there triggers firing? have
> > you traced
>
> > > a session and determined what waits if any are occuring?
>
> > >
>
> > > What expectations do you have for speed?
>
> > >
>
> > > Pascal Ziegler wrote:
>
> > >
>
> > >> hi,
>
> > >> i have to insert several values with a simple insert-
> > statement.
>
> > >> e.g.
>
> > >> INSERT INTO xy VALUES (1, 2, 3, 4); -- really simple ;-)
>
> > >> (the table has just a primary-key index!!)
>
> > >> i did this with oracle 9i and cache from intersystem!
> > cache
>
> > >> was about 5 times faster than oracle!! is this just
>
> > >> the way it is or can i improve the performance of
> > oracle!??
>
> > >> and how?
>
> > >> thanks
>
> > >> - pascal
>
> > >
>
> > >
>
> Use bind variables
>
>
> --
> Posted via http://dbforums.com

Also, think about it, Oracle supports very high rates of transactions - look at tpc-c benchmarks. Do you see cache in the list? (no, wonder why) To measure a system just using single insert as a benchmark is not a very good benchmark. How robust is cache at handling transactions or recovering from a system crash?

Jim Received on Sat Sep 06 2003 - 11:30:04 CDT