slow FreeCell (was: Re: Morphic slow in 2.8 (was:
Forecast))
David Farber
dfarber at numenor.com
Fri Oct 6 16:45:23 UTC 2000
Alright, here is what i came up with. i ran all the tests on a 1024x768
monitor with the squeak window maximised (but not in full screen mode). the
three numbers for each test are for a screen depth of 8, 16, and 32
respectively. i tested 8bit depth because that is what i normally use and
that is what i've run all the FreeCell tests under. the OS display is set
to 32 bits ("True Color").
====Squeak 2.7-1782
"Raw #forceToScreen speed" 14231 17405 6209
"Raw BitBlt speed" 1593 3155 6229
"Raw Balloon speed" 34079 27319 7721
====Squeak 2.9a-2773
"Raw #forceToScreen speed" 14131 17295 6189
"Raw BitBlt speed" 2023 3815 7571
"Raw Balloon speed" 43102 37353 13359
so, it looks like Balloon's speed has taken a hit since 2.7 ...
david
At 09:15 PM 10/5/00 -0700, you wrote:
>David,
>
>Something smells wrong here. Can you do me a favour and check what the
>difference in speed for you is when you run:
>
>"Raw #forceToScreen speed"
> [1 to: 100 do:[:i| Display forceToScreen]] timeToRun.
>
>"Raw BitBlt speed"
> [Display deferUpdates: true.
> 1 to: 100 do:[:i| (FormCanvas on: Display) fillColor: Color blue].
> Display deferUpdates: false] timeToRun.
>
>"Raw Balloon speed"
> [Display deferUpdates: true.
> 1 to: 100 do:[:i|
> (BalloonCanvas on: Display)
> drawRectangle: Display boundingBox
> color: Color red
> borderWidth: 0
> borderColor: nil].
> Display deferUpdates: false] timeToRun.
>
>On my machine (400Mhz PII at 1024x768x16) the above gives:
>Display #forceToScreen BitBlt Balloon
>1024x768x16 4367 1323 10362
>1024x768x32 3318 2449 3884
>
>The difference in speed between 16/32 bit Balloon drawing is due to the
>dithering taking place in 16bit. Beautiful but a little expensive ;-)
>
>[Note to Bob: Could you run the above on your G3 for reference?!]
>
> - Andreas
>
>
--
David Farber
dfarber at numenor.com
More information about the Squeak-dev
mailing list
|