Precisely. Any hints from the real gurus out there as to how that might
be accomplished (or alternatively, reasons why it's hopeless)?
On Wed, 2003-06-18 at 18:42, scott.marlowe wrote:
> It would be nice to have something that could suggest settings for the
> different *cost* options.
>
> On 18 Jun 2003, matt wrote:
>
> > Are there really any performance settings of much interest beyond the
> > shared and non-shared memory settings? Beyond those the interactions
> > get so complex that automation is probably impossible anyway, and
> > certain options like fsync = false should never be 'recommended'.
> >
> > On the other hand, a way of empirically deriving some 'correct'
> > optimizer parameters for a given machine would be very nice :-)
> >
> >
> > Matt
> >
> >
> > On Wed, 2003-06-18 at 18:07, Tom Lane wrote:
> > > Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > > > Sure, it would be great if we could do it.
> > >
> > > If the program actually derives reliable numbers, it would be great.
> > > It could easily do more harm than good if it gives bogus results.
> > > I think it will be very hard to get reliable rather than bogus results
> > > :-( ... but feel free to try.
> > >
> > > regards, tom lane
> > >
> > > ---------------------------(end of broadcast)---------------------------
> > > TIP 6: Have you searched our list archives?
> > >
> > > http://archives.postgresql.org
> > >
> >
> >
> > ---------------------------(end of broadcast)---------------------------
> > TIP 9: the planner will ignore your desire to choose an index scan if your
> > joining column's datatypes do not match
> >
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 8: explain analyze is your friend
>