Re: [GnomeMeeting-devel-list] GnomeMeeting on 2.6

["Damien Sandras" <dsandras seconix com>]

I have the same conclusion as you here. I can reproduce the problem with
a very violent "make -j 2". GnomeMeeting really becomes unusable in
that case.

I wonder if it is related to the preemptive patch or not. I'm trying
without.

I will mail Robert Love about this.

Le 30/1/2004, "Johnny Strom" <jonny strom netikka fi> a écrit:

>Damien Sandras wrote:
>> Hi Johnny,
>>
>> I'm interested in results with Seti home not running. Nothing indicates
>> if Seti home has a good behavior or not.
>
>Well everyting is fine when seti home is not running or when I am not
>compiling somethig. So the sound and the video is fine at the other and
>that is true for bouth settings of P_HAS_SEMAPHORES (1 and 0).
>
>
>What triggers this problem is when something tries to get 100% cpu all
>the time, so for example if something just tries to get 30% cpu then it
>is not a problem like when I am playing two avi movies at the same time,
>that is ok. But when moving some windows around all the time then that
>tries to get 100% cpu and one gets a choppy sound in the other end.
>
>So conclusion this problem is connected to programs that tries to get
>100% cpu it is then when the problem starts to show up.
>
>Cheers Johnny
>
>>
>> What are your communications and responsiveness results on 2.6.1 without
>> SETI and with various settings of P_HAS_SEMAPHORES?
>>
>> Try moving windows and do usual things and see if it has impact over the
>> sound quality. Is the GUI slow only when Seti is eating the CPU ?
>>
>> Le 30/1/2004, "Johnny Strom" <jonny strom netikka fi> a écrit:
>>
>>
>>>Hi
>>>
>>>Now I did a test with P_HAS_SEMAPHORES = 1 it is the same as with 0,
>>>so the sound is bad with seti running but it seems to be a littla better
>>>than when P_HAS_SEMAPHORES was 0. If I run seti with nice level 19 then
>>>it is ok. So now with 2.6.1 kernel and P_HAS_SEMAPHORES = 1 it the best
>>>so far.
>>>
>>>But still it is not like it was in the 2.4 kernel.
>>>
>>>It seems like that GM is punished bye the schedular or something,
>>>the gui becomes reel slow, note that if I try to start other
>>>applications so are the gui fast on the other applications.
>>>
>>>And with seti and GM running so is the load getting high:
>>>load average: 6,75, 3,61, 3,99
>>>
>>>
>>>Could it be becous that GM dose a lot of I/O all the time?
>>>I think there are stuff in the kernel now that looks at that.
>>>
>>>
>>>Cheers Johnny
>>>
>>>
>>>
>>>Johnny Strom wrote:
>>>
>>>>Hi
>>>>
>>>>Now I am testing this with David Quental, so if I am compiling
>>>>at the samt time as GM is running then I get a lot of cuts in the sound
>>>>with P_HAS_SEMAPHORES = 0. This is with 2.6.1 kernel.
>>>>
>>>>
>>>>!Note that without any cpu intensive processes like seti or
>>>>compilation so is the sound fine.
>>>>
>>>>
>>>>I will send an other mail in some minutes where I put
>>>>P_HAS_SEMAPHORES = 1 and repport what happens.
>>>>
>>>>brb
>>>>
>>>>Cheers Johnny
>>>>
>>>>
>>>>Damien Sandras wrote:
>>>>
>>>>
>>>>>Who is using the 2.6 kernel and is ready to test the latest changes with
>>>>>it?
>>>>>
>>>>>Craig, given the fact that many packagers will compile on 2.4, what is
>>>>>your suggestion?
>>>>>
>>>>>I tried on the 2.4 kernel to force the P_HAS_SEMAPHORES to 0 and I have a
>>>>>bad sound quality when compiling, which I had not with P_HAS_SEMAPHORES
>>>>>set to 1.
>>>>>
>>>>>The problem becomes complex. Having a bad sound quality is not
>>>>>problematic. What was problematic on 2.6 was to have people having bad
>>>>>sound quality as soon as they were moving a window.
>>>>>
>>>>>What are your comments?
>>>>>
>>>>>Thank you,
>>>>>
>>>>>
>>>>>Damien
>>>>>_______________________________________________
>>>>>Gnomemeeting-devel-list mailing list
>>>>>Gnomemeeting-devel-list gnome org
>>>>>http://mail.gnome.org/mailman/listinfo/gnomemeeting-devel-list
>>>>>
>>>>
>>>>
>>>_______________________________________________
>>>Gnomemeeting-devel-list mailing list
>>>Gnomemeeting-devel-list gnome org
>>>http://mail.gnome.org/mailman/listinfo/gnomemeeting-devel-list
>>
>>
>>
>