Thread: pgsql: Document random page cost is only 4x seqeuntial, and not 40x.

pgsql: Document random page cost is only 4x seqeuntial, and not 40x.

From
Bruce Momjian
Date:
Document random page cost is only 4x seqeuntial, and not 40x.

Branch
------
master

Details
-------
http://git.postgresql.org/pg/commitdiff/c1d9df4fa227781b31be44a5a3024865a7f48049

Modified Files
--------------
doc/src/sgml/config.sgml |   20 ++++++++++++++++++++
1 files changed, 20 insertions(+), 0 deletions(-)


Re: pgsql: Document random page cost is only 4x seqeuntial, and not 40x.

From
Thom Brown
Date:
On 14 February 2012 21:55, Bruce Momjian <bruce@momjian.us> wrote:
> Document random page cost is only 4x seqeuntial, and not 40x.
>
> Branch
> ------
> master
>
> Details
> -------
> http://git.postgresql.org/pg/commitdiff/c1d9df4fa227781b31be44a5a3024865a7f48049
>
> Modified Files
> --------------
> doc/src/sgml/config.sgml |   20 ++++++++++++++++++++
> 1 files changed, 20 insertions(+), 0 deletions(-)

This change does say that it is 40 times slower:

"The default value can be thought of as modeling random access as 40
times slower than sequential..."

--
Thom

Re: pgsql: Document random page cost is only 4x seqeuntial, and not 40x.

From
Bruce Momjian
Date:
On Tue, Feb 14, 2012 at 10:44:08PM +0000, Thom Brown wrote:
> On 14 February 2012 21:55, Bruce Momjian <bruce@momjian.us> wrote:
> > Document random page cost is only 4x seqeuntial, and not 40x.
> >
> > Branch
> > ------
> > master
> >
> > Details
> > -------
> > http://git.postgresql.org/pg/commitdiff/c1d9df4fa227781b31be44a5a3024865a7f48049
> >
> > Modified Files
> > --------------
> > doc/src/sgml/config.sgml |   20 ++++++++++++++++++++
> > 1 files changed, 20 insertions(+), 0 deletions(-)
>
> This change does say that it is 40 times slower:
>
> "The default value can be thought of as modeling random access as 40
> times slower than sequential..."

I meant to say that it documents why we have a default of 4x, not 40x.
(In fact, real world, it is probably 100x.)

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +

Re: pgsql: Document random page cost is only 4x seqeuntial, and not 40x.

From
Thom Brown
Date:
On 14 February 2012 22:46, Bruce Momjian <bruce@momjian.us> wrote:
> On Tue, Feb 14, 2012 at 10:44:08PM +0000, Thom Brown wrote:
>> On 14 February 2012 21:55, Bruce Momjian <bruce@momjian.us> wrote:
>> > Document random page cost is only 4x seqeuntial, and not 40x.
>> >
>> > Branch
>> > ------
>> > master
>> >
>> > Details
>> > -------
>> > http://git.postgresql.org/pg/commitdiff/c1d9df4fa227781b31be44a5a3024865a7f48049
>> >
>> > Modified Files
>> > --------------
>> > doc/src/sgml/config.sgml |   20 ++++++++++++++++++++
>> > 1 files changed, 20 insertions(+), 0 deletions(-)
>>
>> This change does say that it is 40 times slower:
>>
>> "The default value can be thought of as modeling random access as 40
>> times slower than sequential..."
>
> I meant to say that it documents why we have a default of 4x, not 40x.
> (In fact, real world, it is probably 100x.)

Hmmm.. okay.  I haven't managed to test that myself, but I'll take
your word for it. :)

--
Thom