[BUG][Events] processor time wasted
Stephan Rudlof
sr at evolgo.de
Tue Aug 1 23:15:21 UTC 2000
Tim,
Tim Rowledge wrote:
>
> I just tried using the MessaageTally idea that Dan mentioned nad was
> toldthat with either a browser or a changesorter open that the system
> spends 97% of it's time in Delay>wait.
>
> Whoever it was that reported the original problem will need to give me
> some more information about what windows were open, what was done, how
> the 'waste' time was worked out etc.
>
> tim
>
> --
> Tim Rowledge, tim at sumeru.stanford.edu, http://sumeru.stanford.edu/tim
> Strange OpCodes: RPM: Read Programmer's Mind
Morphic! I have just checked that in MVC there is not such a problem.
OK, some more to the already posted infos below:
- Squeak 2.9alpha #2447,
- Squeak started with flags (-nojit should be superfluous):
squeak -memory 64m -xshm -nojit
- Morphic project,
- switching off browseWithDrag&Drop and smartUpdate doesn't help;
- with EventImageStuff and TPR-EventFix1 changesets filed in;
- VM compiled with last version from SourceForge at Wed Aug 2 00:54:32 MEST
2000,
- SuSE 6.2 Linux,
- sr at Klaus:~/Squeak/work > gcc -v
Reading specs from /usr/lib/gcc-lib/i486-linux/egcs-2.91.66/specs
gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release),
- XFree with fvwm2 window manager and multiple windows,
-> Squeak fills one of them.
Just close all windows in the current project and then open a simple change
sorter by the menu. This leads to a significant processor load at my
machine, with or without the mouse inside the Squeak window. The percentages
are measured by 'top' and ca. ones, of course.
Please ask what you want to know, if this is not sufficient yet.
Greetings,
Stephan
Typical output of top for one simple change sorter open:
12:57am up 1:15, 7 users, load average: 0.57, 0.60, 0.46
67 processes: 65 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 34.3% user, 4.1% system, 0.0% nice, 62.1% idle
Mem: 159496K av, 155420K used, 4076K free, 66380K shrd, 28824K buff
Swap: 136040K av, 388K used, 135652K free 55320K cached
PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
2773 sr 15 0 21008 20M 3964 R 0 30.2 13.1 3:39 squeak
1170 sr 1 0 4356 4356 3292 S 0 2.3 2.7 0:18
LinCVS_0.2.5
1143 sr 1 0 1224 1224 1004 S 0 1.7 0.7 0:26
xosview.bin
1159 sr 0 0 792 792 612 S 0 1.1 0.4 0:12 top
2782 root 2 0 788 788 612 R 0 1.1 0.4 0:00 top
232 root 0 0 24256 23M 4552 S 0 0.9 15.2 1:45 X
1018 sr 0 0 18768 18M 9368 S 0 0.5 11.7 0:20 netscape
235 sr 0 0 1476 1476 980 S 0 0.1 0.9 0:03 fvwm2
...
If top shows below 5% without beeing in the Squeak window, that would be OK
for me (MVC needs less inside, if you don't move the mouse).
Stephan Rudlof wrote:
>
> This bug report applies, if using event mode (with an event VM).
>
> The processor time used by Squeak is about 4%, if the windows are closed
> and there is no activity.
>
> The processor time increases from this 4% to about
>
> 34%, if a simple ChangeSorter window,
> 50%, if a dual ChangeSorter window,
> 57%, if both,
>
> has/have been opened.
>
> This waste of processor time isn't the case for opening a SystemBrowser
> window.
>
> These problems occure with or without applying the TPR-EventFix1
> changeset.
>
> Stephan
>
> P.S.: I'm not actively seeking for this kind of bugs, but my CPU cooler
> starts in my notebook, if there is significant processor activity over
> some time...
>
> --
> Stephan Rudlof (sr at evolgo.de)
> "Genius doesn't work on an assembly line basis.
> You can't simply say, 'Today I will be brilliant.'"
> -- Kirk, "The Ultimate Computer", stardate 4731.3
--
Stephan Rudlof (sr at evolgo.de)
"Genius doesn't work on an assembly line basis.
You can't simply say, 'Today I will be brilliant.'"
-- Kirk, "The Ultimate Computer", stardate 4731.3
More information about the Squeak-dev
mailing list
|