Re: CPU_TIME > ELAPSED_TIME
- From: "Niall Litchfield" <niall.litchfield@xxxxxxxxx>
- To: daniel.fink@xxxxxxxxxxxxxx
- Date: Wed, 24 Oct 2007 08:29:58 +0100
hey, I knew that once...
it occurs to me now though that, perhaps fittingly for a commercial company,
the rule is the exact opposite of the rule for the workers in the vineyard.
I think I can remember that.
On 10/24/07, Daniel Fink <daniel.fink@xxxxxxxxxxxxxx> wrote:
>
> Elapsed time and CPU time are not measured the same. Elapsed time is a
> stop watch. CPU time is a calendar with the rule that you get credit for
> the full day if you are present when the day starts.
>
> You could be on the job for 10 minutes...but get credit for 24 hours if
> those 10 minutes started precisely at midnight. You could also be on the
> job for 23 hours and 50 minutes...and get no credit for hours because
> you were not on the job when the day started.
>
> Cary Millsap covers this in his book "Optimizing Oracle Performance".
>
> --
> Daniel Fink
>
> Oracle Performance, Diagnosis and Training
>
> OptimalDBA http://www.optimaldba.com
> Oracle Blog http://optimaldba.blogspot.com
>
>
>
> Eagle Fan wrote:
> > yes, I also thought about this possibility before.
> >
> > But this query is very simple, it's doing index unique scan on PK
> > index. buffer gets per execution is only 3.
> >
> > I can't image it run on multiple CPU at one time.
> >
> > What do you think?
>
>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
--
Niall Litchfield
Oracle DBA
http://www.orawin.info
Other related posts:
