Re: [GENERAL] _page_cost parameter with values < 1

Thomas Kellerer <spam_eater@gmx.net> writes:
> recently I have seen a Postgres configuration with the following values:
>   seq_page_cost = 0.5
>   random_page_cost = 0.6
> Is there any advantage (or maybe disadvantage) compared to using e.g. 1.0 and 1.2?

That reduces these costs relative to the cpu_xxx_cost ones.  You'd get the
same plans if you scaled *all* the planner cost parameters by the same
amount, but changing only these two is the easiest way to reduce the
significance of I/O relative to CPU costs.

regression=# select name,setting from pg_settings where name like '%cost';
         name         | setting
----------------------+---------
 cpu_index_tuple_cost | 0.005
 cpu_operator_cost    | 0.0025
 cpu_tuple_cost       | 0.01
 parallel_setup_cost  | 1000
 parallel_tuple_cost  | 0.1
 random_page_cost     | 4
 seq_page_cost        | 1
(7 rows)

            regards, tom lane