Re: Oracle speed issue between 8i and 9i

Comment in-line.

--
Regards

Jonathan Lewis
http://www.jlcomp.demon.co.uk

  The educated person is not the person
  who can answer the questions, but the
  person who can question the answers -- T. Schick Jr


One-day tutorials:
http://www.jlcomp.demon.co.uk/tutorial.html
____Belgium__November (EOUG event - "Troubleshooting")
____UK_______December (UKOUG conference - "CBO")


Three-day seminar:
see http://www.jlcomp.demon.co.uk/seminar.html
____UK___November


The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html


"Billy Verreynne" <vslabs_at_onwe.co.za> wrote in message
news:1a75df45.0310220056.3dab15a7_at_posting.google.com...


>

> First, we need to establish a baseline. 9i performance is equal or

> better than 8i ito performance using the same query, same table,

same


> physical db layout, same stats, same platform, etc. Period.

>

> If this someone wants to dispute this, well then they are saying

that


> Oracle's db developers downgraded the performance of 9i in

comparison


> with 8i. And that is bullshit.

>


There is at least one very good reason why the Oracle developers
could choose to make some queries, or even the entire strategy
for handling logical I/O run "more slowly" - and that is to increase
scalability.

I have found several areas of operation where Oracle (the software)
seems to use more CPU to get a task down whilst using less
latching.  Given that CPU speeds keep going up, but latching
reduces scalability, it makes sense to trade CPU for latches,
i.e. to produce slower code that is capable of higher concurrency.

Another line of thought that would also result in a considering
slower code a benefit - Oracle has added a lot more CPU-intensive
instrumentation all the way down the line, but the benefit is that
problems can be diagnosed more readily, and real problems can
be fixed more easily.  If that means current systems run more
slowly, so what - CPUs can be upgraded, it's people-time that
costs most money.