Clément,

There is an optimized route for reprojection from 4326 -> 3857 in MapServer 
(see http://lists.osgeo.org/pipermail/mapserver-dev/2012-February/011956.html) 
so it can be very fast. I can't see how if could be faster than no projection. 
Perhaps the 10% is just within the margin of error? Or did you do the test 
1000's of times and average the results? That might be what it takes to really 
see a difference.

Mike

--
Michael Smith
US Army Corps
Remote Sensing GIS/Center

From: Clément MONIER 
<[email protected]<mailto:[email protected]>>
Date: Tuesday, December 4, 2012 7:23 AM
To: "[email protected]<mailto:[email protected]>" 
<[email protected]<mailto:[email protected]>>
Subject: [mapserver-users] Projection takes less time than no projection ?
Resent-From: Michael Smith 
<[email protected]<mailto:[email protected]>>

Hi All,

I’m using MapServer to generate tiles to overlay on a GoogleMaps base layer.
My data files are WGS84 shapefiles.

So my current mapfile was looking like :

MAP
…
PROJECTION "init=epsg:3785" END
…
LAYER
…
PROJECTION "init=epsg:4326" END
…
END
…
END

I was wondering about improving response time by disabling on-the-fly 
projection.
So I did pre-project my shp into spherical Mercator first.
And my new mapfile now looks like :

MAP
…
PROJECTION    "proj=merc"    "a=6378137"    "b=6378137"    "lat_ts=0.0"    
"lon_0=0.0"    "x_0=0.0"    "y_0=0"    "k=1.0"    "units=m"    "nadgrids=@null" 
   "no_defs" END
…
LAYER
…
END
…
END

I do get the same  tiles.
But when asking a single tile using both mapfile, I noticed that the response 
time was higher (about +10%) when the second mapfile was used !

Do you have any idea about what’s going on ?

Thanks  & Regards

Clément
_______________________________________________
mapserver-users mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/mapserver-users

Reply via email to