Raw graphics performance

radoslav hodnicak rh at 4096.sk
Tue Jun 21 14:13:54 UTC 2005


Huh are you sure you mean milliseconds? Because I get about 4.5 seconds 
(4660ms) on my desktop system in squeak 3.8



On Tue, 21 Jun 2005, Aran Lunzer wrote:

> Hi
>
> Doing some comparisons of PCs as preparation to buying a new laptop, I've
> noticed some unexpectedly large differences in Squeak's graphics
> performance.
>
> I narrowed it down to the speed of DisplayScreen>>forceToScreen:
>
> For example, running the following within an identical 3.7 image on four
> different systems gave the per-loop timings listed below.
>
>   Time millisecondsToRun: [
>     1000 timesRepeat: [ Display forceToScreen: Display boundingBox ] ]
>
> Toshiba dynabook SS LX/190 DK    [laptop]
> 1.6 GHz Pentium M 730
> NVIDIA GeForce Go 6200 TE 64M / 6600 TE 128M
> 6.0ms
>
> NEC LaVie LR 700/9    [laptop]
> 1.6 GHz Pentium
> ATI Mobility Radeon 9600 series
> 8.1ms
>
> Dell Dimension 8200   [large desktop]
> 2 GHz Pentium 4
> NVIDIA GeForce2 GTS
> 19.0ms
>
> Dell Inspiron 8000    [chunky laptop]
> 1 GHz Pentium III
> (not sure about the driver)
> 33.2ms
>
>
> Processor zorch alone doesn't explain it; is the secret in the display
> drivers?  Or something to do with video memory?  Or something that could
> be tweaked on a given computer?
>
> Personally I don't rely on massive display speed, but when working with
> lots of debug output going into the Transcript the difference is certainly
> dramatic  :-)
>
> Many thanks
>
> Aran
> --
> Aran Lunzer
> Meme Media Laboratory
> Hokkaido University
> Sapporo, Japan
>
>
>



More information about the Squeak-dev mailing list