Tom Lane wrote:
>
> Ed Loehr <eloehr@austin.rr.com> writes:
> > What is the status of the genetic algorithm query optimizer? Is this
> > supposed to work well on many-table joins, or has it fallen out of favor
> > or in disrepair?
>
> It's supposed to work ;-). I'm not sure that the default parameters are
> optimal, however. If you experiment with other settings, please post your
> results.
Query time dropped from many minutes to 13 seconds on a 12-table join
with a little tweaking from the default params:
My $PGDATA/pg_geqo:
-------------------
Pool_Size 1024
# Effort high
Generations 100
Random_Seed 330418761
Selection_Bias 2.00
Similar performance with Generations setting of 800 derived from Effort.
Regards,
Ed Loehr
