Re: AW: [HACKERS] Some notes on optimizer cost estimates

Zeugswetter Andreas SB <ZeugswetterA(at)wien(dot)spardat(dot)at> writes:
> My points are:
> 1. even if it is good for an optimizer to be smart,
> it is even more important, that it is predictable

A good point indeed. And unless we find that there is a huge range in
the ratios across different machines, we'd be wasting our time trying to
calibrate the numbers for a particular machine --- we could just as well
use an average value. The optimizer has many other, far worse, sources
of error than that.

> 2. I compile on test machine, production is completely different
> (more processors, faster disks and controllers)

In practice we'd do this at initdb time, not configure time, so I'm
not sure that that's really an issue. But your other point is
well taken.

regards, tom lane