Re: Bitmap Heap scan 8.1/8.2

Martin Marques escribió:
> Martin Marques escribió:
>> Pavel Stehule wrote:
>>>
>>> try
>>>
>>> set work_mem to '8MB';
>>> and
>>> explain analyze select ..
>> These things didn't help. What changed the plan completely was this:
>> seq_page_cost = 5.0 # measured on an arbitrary scale
>> cpu_tuple_cost = 0.05 # same scale as above
>
> Can someone explain how this parameters are measured? What is 5.0 in this
> case for seq_page_cost?

It's an arbitrary number, based on which all the other numbers are
measured.

What people generally do around here is mess with random_page_cost, and
leave seq_page_cost alone. Often, it's the ratio
seq_page_cost/random_page_cost what's most important to the cost
equations results. (seq_page_cost wasn't tunable at all until
recently, say 8.1 or 8.2 AFAIR).

--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.