Re: The New GNUstep Seems Slow

[Eric Wasylishen]

I'd suggest trying kcachegrind (nice gui profiler from what I recall) or gprof.

For gprof you have to do a 'make clean' of everything then 'make profile=yes' - 
I think kcachegrind works on unmodified builds.

I'll give these a try when I have a chance - sometime GS feels a bit slow for 
me too.

Eric

On 2011-04-28, at 4:08 AM, Fred Kiefr wrote:

> I don't know about any specific reason why GNUstep should now be slower. This 
> seems to be an important issue to investigate. Which bacend are you using? A 
> wrong backend is the most common reason for a slowdown. If this isn't the 
> case we need to use tools to find out where the time gets spend. I will send 
> a mail on this next week, when I am back home.
> 
> Fred
> 
> On the road
> 
> Am 27.04.2011 um 07:07 schrieb Germán Arias <german@xelalug.org>:
> 
>> Yes, I noticed too that the new GNUstep is a bit slow. But not too. On
>> my machine, GWorkspace works fine and fast. So your problem should be
>> something with configuration or installation.
>> 
>> 
>> On mar, 2011-04-26 at 18:12 +0100, Richard Stonehouse wrote:
>>> GNUstep built from the recent tarballs:
>>> 
>>>   gnustep-make-2.6.0
>>>   gnustep-base-1.22.0
>>>   gnustep-gui-0.20.0
>>>   gnustep-back-0.20.0
>>> 
>>> runs but seems very slow. On launching GWorkspace, it takes approx
>>> 30 - 35 secs before a blank window appears, and a further 10 - 15
>>> secs before this gets filled in with the file browser display. During
>>> the whole of this time GWorkspace is taking nearly 100% of the CPU. In
>>> the previous version (make-2.4.0, base-1.20.1, gui- and back-0.18.0)
>>> the whole sequence used to take just 2 - 3 secs.
>>> 
>>> Other operations in GWorkspace, e.g. moving to an adjacent column in
>>> the display, are also slow and CPU-intensive. Other applications,
>>> e.g. SystemPreferences, show similar but less extreme symptoms.
>>> 
>>> It may well be that I've made an error in the build, but the only
>>> obviously suspicious thing is a message in the gnustep-base build
>>> output:
>>> 
>>>   "gnustep-base-1.22.0-1130.1-results.txt:checking for thread-safe
>>>   +initialize in runtime... configure: WARNING: Your ObjectiveC
>>>   runtime does not support thread-safe class initialisation.  Please
>>>   use a different runtime if you intend to use threads."
>>> 
>>> The machine is single-processor and the Objective C library is
>>> 
>>>   libobjc45-4.5.0_20100604
>>> 
>>> from the openSUSE 11.3 distribution.
>>> 
>>> Is this a known problem? (I seem to remember some discussion of
>>> diagnostic code slowing things down but assume this has been removed
>>> in the tarball release).
>>> 
>>> If not, what further diagnostics would be useful?
>>> 
>> 
>> 
>> _______________________________________________
>> Discuss-gnustep mailing list
>> Discuss-gnustep@gnu.org
>> https://lists.gnu.org/mailman/listinfo/discuss-gnustep
> 
> _______________________________________________
> Discuss-gnustep mailing list
> Discuss-gnustep@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep