GD vs AGG performance...

[John Cole]

Jeff Hoffmann wrote:
> 
> The AGG image is about 4x the file size as the GD image, which isn't 
> that surprising with all that antialiasing going on.  More of a symptom 
> than anything, but not everyone is going to enjoy downloading 400k 
> images, so it's something to think about if you're going to use this in 
> a production site.
> 

Your right there, I switched it to JPG and it's 70KB.  Previously, PNG
sharpness over JPG made the small file size difference moot, but with all of
these smoothed features, the better compression of JPG seems to be better.


Jeff Hoffmann wrote:
> 
> I think the others are right, it's the shear number of features you're 
> drawing (it takes 0.5 sec just for your street layer) -- lot of 
> smoothing = more time + bigger PNG.
> 

When you say 'shear number', I'm thinking this is a small dataset :-) 
Though the size of the output image shouldn't effect render time as much as
compression time (and JPG is just slightly slower than PNG)/

The crux of my question is; does AGG performance degrade as much as I'm
seeing, or are there other factors in MS itself that are causing this (eg.
debug flags still on, known optimizations not yet performed, etc.)  From
what I'm hearing, it's a new feature, and performance hasn't really been
addressed.  Till then, I'll try and restrict it's use to TileCache backends
or smaller datasets.  But dang, it looks soooo good.

John



-- 
View this message in context: http://www.nabble.com/GD-vs-AGG-performance...-tf4319265.html#a12302079
Sent from the Mapserver - User mailing list archive at Nabble.com.