[Benchmarking] Raster and vector template scripts available

Andrea Aime aaime at opengeo.org
Fri Aug 13 09:35:54 EDT 2010

Martin Daly ha scritto:
>> And well, let me know how the scripts work for you. I'm completely 
>> wiped out for the day, will look more into the scripts tomorrow.
> I've generated all 2200 EPSG:4326 vector images and will trawl
> through looking for blanks when I get a chance. I can double check
> them against other servers just in case, and then generate (hopefully
> only a handful) of replacement CSV rows using the script.

Sure, that can be good. It would be better to have the same on the
raster data series where the number of blank tiles (or black tiles
if you have such background) are much much more. Well, enough
that we should look for an algorithmic solution to the problem...

> I'm going to be a massive PITA and suggest that the 3875 requests
> might be more appropriate if they corresponded exactly to Google Maps
> etc. tile size and extents (at two or three different zoom levels).
> Isn't this the more likely use case, overlaying WMS on top of GMaps
> and so on?

That would a completely different test, we'd loose any ability
to make even the slight parallel with the non reprojected benchmark.
The benchmarks are setup exactly to make it impossible to cache
the results, whilst what you suggest is designed to be cacheable.

-1 from me, let's see how the others feel like about this change.


Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.

More information about the Benchmarking mailing list