Re: How is random_page_cost=4 ok? - Mailing list pgsql-hackers

From Greg Smith
Subject Re: How is random_page_cost=4 ok?
Date
Msg-id Pine.GSO.4.64.0810101347200.204@westnet.com
Whole thread Raw
In response to Re: How is random_page_cost=4 ok?  (Gregory Stark <stark@enterprisedb.com>)
Responses Re: How is random_page_cost=4 ok?
Re: How is random_page_cost=4 ok?
Re: How is random_page_cost=4 ok?
List pgsql-hackers
On Fri, 10 Oct 2008, Gregory Stark wrote:

> They don't quote sustained bandwidth for consumer drives but 50-60MB/s are the
> numbers I remembered -- admittedly from more than a couple years ago. I didn't
> realize 7200 RPM drives had reached such speeds yet.

The cheap ($42!) 7200RPM SATA disks I bought a stack of for my home server 
hit a sequential 110MB/s at the beginning edge, at the other end 
throughput is still 60-70MB/s.  The smaller capacities of Seagate's 
7200.11 average about 100MB/s nowadays.  But by the time you seek to a 
location (8-9ms) and line the heads up (half a rotation at 7200RPM 
averages 4ms) you can easily end up at 12-13ms or higher measured access 
time on random reads with those.  So the true random/sequential ratio 
reaches crazy numbers.

I don't think random_page_cost actually corresponds with any real number 
anymore.  I just treat it as an uncalibrated knob you can turn and 
benchmark the results at.

--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD


pgsql-hackers by date:

Previous
From: Michael Renner
Date:
Subject: Re: How is random_page_cost=4 ok?
Next
From: Emmanuel Cecchet
Date:
Subject: Re: Transactions and temp tables