Investigating fullBounds/layoutChanged

Eddie Cottongim squeak-dev at lists.squeakfoundation.org
Fri Oct 18 16:43:04 UTC 2002


5-10 collapsed SystemWindows, a trash can, and an open Browser or two.

On Windows(mayber others too) mouse events come in at a fixed rate (35hz ?),
so it would take a smaller amount of time for faster processors to handle
all the events, while a slow enough processor might spend all its time
handling them. That roughly fits my two measurements of 42%@133mhz and
8%@600mhz ( one quarter the time used in a system 4x faster ).

A fast enough machine would spend so little time doing mouse moves that it
wouldn't show up at all in the time profile. I'm suprised that 700 mhz is
fast enough to cause that, though.

Eddie

----- Original Message -----
From: "Bob Arning" <arning at charm.net>
To: <squeak-dev at lists.squeakfoundation.org>
Sent: Friday, October 18, 2002 7:52 AM
Subject: Re: Investigating fullBounds/layoutChanged


> Wow! What do you have in your World? I asked because I was unable (on a
700 MHz system) to get either to appear at all. My twirling-the-mouse tally:
>
>  - 293 tallies, 4877 msec.
>
> **Tree**
> 100.0% {4877ms} PasteUpMorph>>doOneCycle
>   100.0% {4877ms} WorldState>>doOneCycleFor:
>     89.1% {4345ms} WorldState>>interCyclePause:
>       |89.1% {4345ms} Delay>>wait
>     10.9% {532ms} WorldState>>doOneCycleNowFor:
>       10.9% {532ms} HandMorph>>processEvents
>         10.6% {517ms} EventSensor>>nextEvent
>           10.6% {517ms} EventSensor>>nextEventFromQueue
>
> Cheers,
> Bob






More information about the Squeak-dev mailing list