Scripting performances and PS versions

[undavide]

Hello,
I've noticed some strange behaviors - in terms of PS performances running scripts: if someone more expert than me would like to throw in some hypothesis, I'll be glad to read them.

1) PS CS4 completes a script routine 3 to 4 times faster than the very same routine launched from a panel built with Flash Builder.

2) PS CS5 speeds (panel vs. script) are equal if the panel is built with Flash Builder.

3) Not if the panel is built with Configurator (I'm waiting for confirmation from a friend who tested this one)

The last consideration is unrelated to scripts, nevertheless left me speechless. To test this I made a script who flattens, runs 3 filters (Maximum, Minimum, GaussianBlur):

PS CS5@64bits is faster processing 16bit images than CS5@32bit (1/5 faster)
BUT:
with 8bit images the opposite is true, so CS5@32bit is faster than CS5@64bit (4 to 5 times faster )
How can this be, I don't really know.

Any comment is welcome!
Thank you,

Davide

[cmyk]

This article talks about CS5's inability to allocate memory correctly when running a script for the first time on launch of the application. It may not speak to your problems directly but does make some interesting points.

[undavide]

cmyk wrote:This article talks about CS5's inability to allocate memory correctly when running a script
Thanks for the link, I read it some time ago and it's still surprising
Nevertheless the CS5 speed difference has been tested restarting the app each time... can't really figure out why such a difference from 64/32 bit dealing with 8/16 bits files.
Regards,

Davide