[postgis-users] strange benchmark results

[Andreas Brandl]

Hi Mark,

------ "Mark Cave-Ayland" <mark.cave-ayland at siriusit.co.uk> schrieb:
> Andreas wrote:
> > [...]
> > The last days I ran the benchmark for PostGIS and had a first look
> on the results. Some test cases show a strange behaviour which I like
> to share here hoping for some feedback.
> > 
> > [...]
> 
 
> You should find that the results should be the same across any
> platform; 

'platform' meaning either OS (i.e. win vs. linux) or architecture (i.e. 32 vs. 64-bit) or both (in this context)?

> The first questions I would like to ask are:
> 
> i) Do you re-open a connection for every SELECT command you issue to
> the 
> database, or do you use a new connection? 

I do always use the same Connection for each and every test. This is mainly because this way it is more like a real world scenario in my opinion. 

> If you don't, does it make a 
> difference if you do?

I will test with reconnecting for each test. However this will take some time and I'm going to post results as soon as I get them.

> ii) Which versions of PostgreSQL/PostGIS are you using on both Windows
> 
> and Ubuntu? The output of 'SELECT version(), postgis_full_version()' 
> from both setups would be useful. In particular, this will confirm
> which 
> versions of GEOS are being used in each setup.

The output on the windows machine (i.e. were those strange results originated from) is:

PostgreSQL 8.4.1, compiled by Visual C++ build 1400, 32-bit 
POSTGIS="1.4.0" GEOS="3.1.1-CAPI-1.6.0" PROJ="Rel. 4.6.1, 21 August 2008" USE_STATS

The postgres installation on ubuntu:

PostgreSQL 8.4.1 on x86_64-unknown-linux-gnu, compiled by GCC gcc (Ubuntu 4.4.1-4ubuntu8) 4.4.1, 64-bit
POSTGIS="1.4.0" GEOS="3.1.1-CAPI-1.6.0" PROJ="Rel. 4.7.1, 23 September 2009" USE_STATS

Apart from proj, architecture and compiler this looks quite the same.

Thanks for your reply, this is a really interesting topic to me. :)

Best Regards,
Andreas