[postgis-users] Effect of tile size on raster query times

[Duncan Golicher]

Following a discussion with Pierre and Bborie I have posted the results of
running some quick tests on the influence of tile size and storage type on
query run times.

http://duncanjg.wordpress.com/2013/09/21/effect-of-tile-size-and-data-storage-on-postgis-raster-query-times/

The tests can be easily replicated by adapting the R code (run under Linux
using system commands) in order to compare results on different systems and
with different data.

I hope this might help other users when selecting  tile size..

The optimum for polygon overlays will depend to some degree on the size and
complexity of the polygons with respect to the resolution of the raster.
The relationship was not very sensitive over a range of 200 to 400 pixels
in the particular example used. The same range was also optimum for point
overlays, providing the data was held out of the database. If these results
hold with  other examples (and they seem to do in my experience) then the
advice to use tile sizes of 255 pixels appears to be good.

However. When data are held in the database point on raster queries are
slower. In the case of point overlays they may be orders of magnitude
slower and there seems to be a linear relationship with tile size (smaller
the better). This is awkward, as small tiles are slower to load and the
optimum for point overalys does not coincide with that for polygon overlays
(which seem to take around twice as much time as they do with out db data).

I ran the tests using the latest development version. Tips for compiling on
Ubuntu are here.

http://duncanjg.wordpress.com/2013/09/21/keeping-postgis-up-to-date-compiling-from-source-and-updating-databases/

Duncan




-- 
Dr Duncan Golicher
Investigador Titular,
El Colegio de la Frontera Sur, Chiapas,Mexico
Mexico tel +52 1 967 137 94 20
Skype name duncangolicher

Publications: http://www.mendeley.com/profiles/duncan-golicher

Senior lecturer, Bournemouth University, UK
Centre for Conservation Ecology & Environmental Change
School of Applied Sciences
Christchurch House rm C218a
Bournemouth University
Fern Barrow
Poole (Dorset) BH12 5BB UK
Tel. +44 (0)1202 961682

For list of publications see Researcher ID:
http://www.researcherid.com/rid/B-4240-2009

dgolicher at bournemouth.ac.uk
dgoliche at ecosur.mx

Researcher ID:
http://www.researcherid.com/rid/B-4240-2009
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20130922/5fd66763/attachment.html>