1= In all these results I'm seeing, no one has yet reported what
their physical IO subsystem is... ...when we are benching a DB.
2= So far we've got ~ a factor of 4 performance difference between
Michael Stone's 1S 1C Netburst era 2.5GHz P4 PC and Guido Neitzer's
1S 2C MacBook Pro 2.33GHz C2D. If the physical IO subsystems are
even close to equivalent across the systems benched so far, we've
clearly established that pg performance is more sensitive to factors
outside the physical IO subsystem than might usually be thought with
regard to a DBMS. (At least for this benchmark SW.)
3= Daniel van Ham Colchete is running Gentoo. That means every SW
component on his box has been compiled to be optimized for the HW it
is running on.
There may be a combination of effects going on for him that others
not running a system optimized from the ground up for its HW do not see.
4= If we are testing arch specific compiler options and only arch
specific compiler options, we should remove the OS as a variable.
Since Daniel has presented evidence in support of his hypothesis, the
first step should be to duplicate his environment as =exactly= as
possible and see if someone can independently reproduce the results
when the only significant difference is the human involved. This
will guard against procedural error in the experiment.
Possible Outcomes
A= Daniel made a procedural error. We all learn what is and to avoid it.
B= The Gentoo results are confirmed but no other OS shows this
effect. Much digging ensues ;-)
C= Daniel's results are confirmed as platform independent once we
take all factor into account properly
We all learn more re: how to best set up pg for highest performance.
Ron Peacetree
At 01:35 AM 12/12/2006, Greg Smith wrote:
>On Mon, 11 Dec 2006, Michael Stone wrote:
>
>>Can anyone else reproduce these results? I'm on similar hardware
>>(2.5GHz P4, 1.5G RAM)...
>
>There are two likely candidates for why Daniel's P4 3.0GHz
>significantly outperforms your 2.5GHz system.
>
>1) Most 2.5GHZ P4 processors use a 533MHz front-side bus (FSB); most
>3.0GHZ ones use an 800MHz bus.
>
>2) A typical motherboard paired with a 2.5GHz era processor will
>have a single-channel memory interface; a typical 3.0GHZ era board
>supports dual-channel DDR.
>
>These changes could easily explain the magnitude of difference in
>results you're seeing, expecially when combined with a 20% greater
>raw CPU clock.
