Re: WINE + Galciv + 2.6.0-test3-mm1-O15

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Voluspa]

On 2003-08-12 14:42:02 gaxt wrote:

[...]
> Galciv plays videos quite smoothly but as soon as I run it it will
> freeze the cursor for 12-15 seconds every half-minute or so even
> within the game itself which is turn-based strategy without a lot of
> whizbang stuff. In the past, the videos would stutter but the game
> would not suffer from more than short pauses now and then.

Similar experience here running the game-test (xfree86 4.3.99.10, winex
3.1 and "Baldurs Gate I") on a PII 400 with 128 meg ram. Using
2.6.0-test3-O14.1int + O14.1-O15int.

I would say this is the best _and_ worst scheduler I've tried since Con
had to adopt his work to the new Ingo infrastructure. Slightly
smoother than pure A3 unless you manage to trigger badness. Then you (or
I, at least) get exactly 15 seconds of total freezes, mixed with
different periods of normality. And by freezes, I mean total freezes.

I can switch to an already opened text console, but there
a running "top" is dead and it won't take any keystrokes until the
freeze has gone. The only thing not dead is the "sound repeats" that
Daniel Phillips explained in:

http://marc.theaimsgroup.com/?l=linux-kernel&m=105966612027670&w=2

--quote--
To convince yourself of this, note that when DMA refill fails to meet
its deadline you will hear repeats, not skipping, because the DMA
hardware on the sound card has been set up to automatically restart the
DMA each time the buffer expires.  Try running the kernel under kgdb and
breaking to the monitor while sound is playing.
--unquote--

Mvh
Mats Johannesson

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Con Kolivas]

On Wed, 13 Aug 2003 01:23, Voluspa wrote:
> On 2003-08-12 14:42:02 gaxt wrote:
> Similar experience here running the game-test (xfree86 4.3.99.10, winex
> 3.1 and "Baldurs Gate I") on a PII 400 with 128 meg ram. Using
> 2.6.0-test3-O14.1int + O14.1-O15int.

Yes known issue for reason I mentioned. Currently investigating.

> I would say this is the best _and_ worst scheduler I've tried since Con

Can you give us some idea what your machine is like when it's not running a 
cpu hog win32 game in wine? Also can you try running your game nice +1

Con

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Voluspa]

On Wed, 13 Aug 2003 07:15:22 +1000
Con Kolivas wrote:

> On Wed, 13 Aug 2003 01:23, Voluspa wrote:
> > On 2003-08-12 14:42:02 gaxt wrote:
> > Similar experience here running the game-test (xfree86 4.3.99.10,
> > winex 3.1 and "Baldurs Gate I") on a PII 400 with 128 meg ram. Using
> > 2.6.0-test3-O14.1int + O14.1-O15int.
> 
> Yes known issue for reason I mentioned. Currently investigating.

Yupp, you posted while I was writing.

> > I would say this is the best _and_ worst scheduler I've tried since
> > Con
> 
> Can you give us some idea what your machine is like when it's not
> running a cpu hog win32 game in wine? Also can you try running your
> game nice +1

As I've said, it is extremely hard for me to notice the scheduler unless
I run something graphic intensive, but my comments are of course not
solely based on that game-test. A general feel of smoothness can always
be attributed to the placebo effect, though I challenge anyone to
dispute the following experience.

Fourteen days ago I took home a precompiled Blender 2.28 - previous
copy of this 3D render program to reside here was 2.23. Grabbing the
grid which constitutes the world plane (hold third mouse button, drag)
and rotating it around the axes, I noticed something not present in the
old version. Jerks...

In kernel time this is 2.6.0-test1/2 and Cons O11int. Since then I've
done the "world rotate" on plain kernels and all O-patches without
noticing any improvement. Either mouse pointer or grid gets stuck for
about 1/2 to 2 seconds, making the movement unsynchronized. Can happen
while moving slowly, or spinning like crazy.

I don't have DRI with my mach64 (8 meg) because of choice, use Xv
instead, so I had made up my mind about the blame. Sloppy programming
and weak CPU.

Until O15int enters the mix. Perfect sync all the time, no matter how
slow, fast or long I rotate the world plane. That is why I call this
scheduler the "best". The "worst" part comes from the "known issue" but
also because I happened to run xmms on a directory of mp3s while
rotating. I got severe blackouts in the music, ca 10 - 15 seconds long.
When the blackout starts I no longer have to move the mouse, it is
enough to hold down the button. The second I release it, the music
returns.

Xmms has not been involved in the rotating before, so can't tell if the
blackouts have occurred. Could be due to the "known issue", but I'll
retest with older kernels when time permits.

Running the game-test with nice +1... Well, I tried that first on
wineserver, then wine, then X, after which I put -1 on them
consecutively. No real pattern emerged. Badness all around. And now I'm
falling asleep. Bye.

Mvh
Mats Johannesson

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Timothy Miller]

Voluspa wrote:

> When the blackout starts I no longer have to move the mouse, it is
> enough to hold down the button. The second I release it, the music
> returns.


I think this sort of thing has been discussed before.  I get the 
impression that xmms blocks on the X server, so when some app grabs the 
server, then xmms gets blocked and stops.  I don't know why the display 
code is not in a separate thread from the audio code; although maybe 
they are but they interact somehow that causes this.

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Con Kolivas]

On Thu, 14 Aug 2003 04:45, Timothy Miller wrote:
> Voluspa wrote:
> > When the blackout starts I no longer have to move the mouse, it is
> > enough to hold down the button. The second I release it, the music
> > returns.
>
> I think this sort of thing has been discussed before.  I get the
> impression that xmms blocks on the X server, so when some app grabs the
> server, then xmms gets blocked and stops.  I don't know why the display
> code is not in a separate thread from the audio code; although maybe
> they are but they interact somehow that causes this.

This is a pure sheduler starvation issue which I'm trying to fix.

Con

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Mike Galbraith]

[-- Attachment #1: Type: text/plain, Size: 2726 bytes --]

At 01:34 PM 8/13/2003 +1000, Con Kolivas wrote:
>On Wed, 13 Aug 2003 04:24, Mike Galbraith wrote:
> > At 12:40 AM 8/13/2003 +1000, Con Kolivas wrote:
> > >On Wed, 13 Aug 2003 00:42, gaxt wrote:
> > > > Photoshop 6 (yes, legal owned version) in wine is flawless (as it was
> > > > with 2.6.0-test3)
> > > >
> > > > Galciv plays videos quite smoothly but as soon as I run it it will
> > > > freeze the cursor for 12-15 seconds every half-minute or so even within
> > > > the game itself which is turn-based strategy without a lot of whizbang
> > > > stuff. In the past, the videos would stutter but the game would not
> > > > suffer from more than short pauses now and then.
> > >
> > >Yes, herein lies one of those mysteries that still eludes me but I have
> > > been investigating it. I can now reproduce in other applications what
> > > appears to be the problem - Two cpu hogs, X and evolution for example are
> > > running and evolution is making X the cpu hog. The problem is that X gets
> > > demoted whereas evolution doesn't. Strangely, dropping evolution to nice
> > > +1 or making X -1 seems to change which one gets demoted, and X is now
> > > much smoother. I assume the same thing is happening here between wine and
> > > wineserver, which is why you've seen reversal of priorities in your
> > > previous posts. See if renicing one of them +1 helps for the time being.
> > > I will continue investigating to find out why the heck this happens and
> > > try and fix it.
> > >
> > >Con
> > >
> > >P.S. I've cc'ed MG because he has seen the scheduler do other forms of
> > >trickery and may have thoughts on why this happens.
> >
> > That sounds suspiciously similar to my scenario, but mine requires a third
> > element to trigger.
> >
> > <scritch scritch scritch>
> >
> > What about this?  In both your senario and mine, X is running low on cash
> > while doing work at the request of a client right?  Charge for it.  If X is
> > lower on cash than the guy he's working for, pick the client's pocket...
> > take the remainder of your slice from his sleep_avg for your trouble.  If
> > you're not in_interrupt(), nothing's free.  Similar to Robinhood, but you
> > take from the rich, and keep it :)  He's probably going straight to the
> > bank after he wakes you anyway, so he likely won't even miss it.  Instead
> > of backboost of overflow, which can cause nasty problems, you could try
> > backtheft.
>
>Not a bad idea at all. The working for someone else thing is killing me. Now,
>how to implement...

I had to back up and regroup a bit because of backboost sanity problems 
(wish I could pull those dang fangs, backboost is wonderful otherwise), but 
the attached cured my inversion problem.

         -Mike 

[-- Attachment #2: xx.diff --]
[-- Type: application/octet-stream, Size: 1400 bytes --]

--- linux-2.6.0-test1.G8/kernel/sched.c.org	Wed Aug 13 09:19:13 2003
+++ linux-2.6.0-test1.G8/kernel/sched.c	Wed Aug 13 14:41:42 2003
@@ -358,6 +358,8 @@
 
 	if (sleep_time > 0) {
 		unsigned long long sleep_avg;
+		unsigned long run_time = now - current->timestamp;
+		unsigned int slice = 1000000 * current->time_slice;
 
 		/*
 		 * This code gives a bonus to interactive tasks.
@@ -381,6 +383,22 @@
 			p->sleep_avg = sleep_avg;
 			p->prio = effective_prio(p);
 		}
+		/*
+		 * If the awakened task has been asleep for longer than the
+		 * waker has had the CPU plus round-robin time, and it is
+		 * going to preempt, there is a good chance that the waker
+		 * is not getting enough CPU to service the awakened task in
+		 * a timely manner, and that this is the cause of the preempt.
+		 * Take some of the resulting sleep_time from the awakened
+		 * task, and give it to the waker.
+		 */
+		if (!in_interrupt() && p->mm && current->mm && sleep_avg >
+				current->sleep_avg + slice && sleep_time > run_time +
+				(1000000 * this_rq()->nr_running * TIMESLICE_GRANULARITY) &&
+				TASK_PREEMPTS_CURR(p, task_rq(current))) {
+			current->sleep_avg += slice;
+			sleep_avg -= slice;
+		}
 	}
 }
 
@@ -1414,6 +1432,7 @@
 		next->timestamp = now;
 		rq->nr_switches++;
 		rq->curr = next;
+		next->timestamp = now;
 
 		prepare_arch_switch(rq, next);
 		prev = context_switch(rq, prev, next);

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Con Kolivas]

On Wed, 13 Aug 2003 04:24, Mike Galbraith wrote:
> At 12:40 AM 8/13/2003 +1000, Con Kolivas wrote:
> >On Wed, 13 Aug 2003 00:42, gaxt wrote:
> > > Photoshop 6 (yes, legal owned version) in wine is flawless (as it was
> > > with 2.6.0-test3)
> > >
> > > Galciv plays videos quite smoothly but as soon as I run it it will
> > > freeze the cursor for 12-15 seconds every half-minute or so even within
> > > the game itself which is turn-based strategy without a lot of whizbang
> > > stuff. In the past, the videos would stutter but the game would not
> > > suffer from more than short pauses now and then.
> >
> >Yes, herein lies one of those mysteries that still eludes me but I have
> > been investigating it. I can now reproduce in other applications what
> > appears to be the problem - Two cpu hogs, X and evolution for example are
> > running and evolution is making X the cpu hog. The problem is that X gets
> > demoted whereas evolution doesn't. Strangely, dropping evolution to nice
> > +1 or making X -1 seems to change which one gets demoted, and X is now
> > much smoother. I assume the same thing is happening here between wine and
> > wineserver, which is why you've seen reversal of priorities in your
> > previous posts. See if renicing one of them +1 helps for the time being.
> > I will continue investigating to find out why the heck this happens and
> > try and fix it.
> >
> >Con
> >
> >P.S. I've cc'ed MG because he has seen the scheduler do other forms of
> >trickery and may have thoughts on why this happens.
>
> That sounds suspiciously similar to my scenario, but mine requires a third
> element to trigger.
>
> <scritch scritch scritch>
>
> What about this?  In both your senario and mine, X is running low on cash
> while doing work at the request of a client right?  Charge for it.  If X is
> lower on cash than the guy he's working for, pick the client's pocket...
> take the remainder of your slice from his sleep_avg for your trouble.  If
> you're not in_interrupt(), nothing's free.  Similar to Robinhood, but you
> take from the rich, and keep it :)  He's probably going straight to the
> bank after he wakes you anyway, so he likely won't even miss it.  Instead
> of backboost of overflow, which can cause nasty problems, you could try
> backtheft.

Not a bad idea at all. The working for someone else thing is killing me. Now, 
how to implement...

Con

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Mike Galbraith]

At 02:44 PM 8/12/2003 -0400, Timothy Miller wrote:


>Mike Galbraith wrote:
>
>>That sounds suspiciously similar to my scenario, but mine requires a 
>>third element to trigger.
>><scritch scritch scritch>
>>What about this?  In both your senario and mine, X is running low on cash 
>>while doing work at the request of a client right?  Charge for it.
>>If X is lower on cash than the guy he's working for, pick the client's 
>>pocket... take the remainder of your slice from his sleep_avg for your 
>>trouble.  If you're not in_interrupt(), nothing's free.  Similar to 
>>Robinhood, but you take from the rich, and keep it :)  He's probably 
>>going straight to the bank after he wakes you anyway, so he likely won't 
>>even miss it.  Instead of backboost of overflow, which can cause nasty 
>>problems, you could try backtheft.
>
>
>How is this different from back-boost?

With backboost, you take everything that overflows MAX_SLEEP_AVG and give 
it all to the waker... you always pull-up.  With back-theft (blech;), 
there's constant pull-up and push-down for all parties instead of only 
those who reach MAX_SLEEP_AVG, so while you'll still tend to group tasks 
which are related (the original goal of backboost), it shouldn't (wild 
theory) go raging out of control.

         -Mike 

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Timothy Miller]

Mike Galbraith wrote:

> 
> That sounds suspiciously similar to my scenario, but mine requires a 
> third element to trigger.
> 
> <scritch scritch scritch>
> 
> What about this?  In both your senario and mine, X is running low on 
> cash while doing work at the request of a client right?  Charge for it.  
> If X is lower on cash than the guy he's working for, pick the client's 
> pocket... take the remainder of your slice from his sleep_avg for your 
> trouble.  If you're not in_interrupt(), nothing's free.  Similar to 
> Robinhood, but you take from the rich, and keep it :)  He's probably 
> going straight to the bank after he wakes you anyway, so he likely won't 
> even miss it.  Instead of backboost of overflow, which can cause nasty 
> problems, you could try backtheft.


How is this different from back-boost?

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Mike Galbraith]

At 12:40 AM 8/13/2003 +1000, Con Kolivas wrote:
>On Wed, 13 Aug 2003 00:42, gaxt wrote:
> > Photoshop 6 (yes, legal owned version) in wine is flawless (as it was
> > with 2.6.0-test3)
> >
> > Galciv plays videos quite smoothly but as soon as I run it it will
> > freeze the cursor for 12-15 seconds every half-minute or so even within
> > the game itself which is turn-based strategy without a lot of whizbang
> > stuff. In the past, the videos would stutter but the game would not
> > suffer from more than short pauses now and then.
>
>Yes, herein lies one of those mysteries that still eludes me but I have been
>investigating it. I can now reproduce in other applications what appears to
>be the problem - Two cpu hogs, X and evolution for example are running and
>evolution is making X the cpu hog. The problem is that X gets demoted whereas
>evolution doesn't. Strangely, dropping evolution to nice +1 or making X -1
>seems to change which one gets demoted, and X is now much smoother. I assume
>the same thing is happening here between wine and wineserver, which is why
>you've seen reversal of priorities in your previous posts. See if renicing
>one of them +1 helps for the time being. I will continue investigating to
>find out why the heck this happens and try and fix it.
>
>Con
>
>P.S. I've cc'ed MG because he has seen the scheduler do other forms of
>trickery and may have thoughts on why this happens.

That sounds suspiciously similar to my scenario, but mine requires a third 
element to trigger.

<scritch scritch scritch>

What about this?  In both your senario and mine, X is running low on cash 
while doing work at the request of a client right?  Charge for it.  If X is 
lower on cash than the guy he's working for, pick the client's pocket... 
take the remainder of your slice from his sleep_avg for your trouble.  If 
you're not in_interrupt(), nothing's free.  Similar to Robinhood, but you 
take from the rich, and keep it :)  He's probably going straight to the 
bank after he wakes you anyway, so he likely won't even miss it.  Instead 
of backboost of overflow, which can cause nasty problems, you could try 
backtheft.

         -Mike 

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[gaxt]

Photoshop 6 (yes, legal owned version) in wine is flawless (as it was 
with 2.6.0-test3)

Galciv plays videos quite smoothly but as soon as I run it it will 
freeze the cursor for 12-15 seconds every half-minute or so even within 
the game itself which is turn-based strategy without a lot of whizbang 
stuff. In the past, the videos would stutter but the game would not 
suffer from more than short pauses now and then.

Re: WINE + Galciv + 2.6.0-test3-mm1-O15

[Con Kolivas]

On Wed, 13 Aug 2003 00:42, gaxt wrote:
> Photoshop 6 (yes, legal owned version) in wine is flawless (as it was
> with 2.6.0-test3)
>
> Galciv plays videos quite smoothly but as soon as I run it it will
> freeze the cursor for 12-15 seconds every half-minute or so even within
> the game itself which is turn-based strategy without a lot of whizbang
> stuff. In the past, the videos would stutter but the game would not
> suffer from more than short pauses now and then.

Yes, herein lies one of those mysteries that still eludes me but I have been 
investigating it. I can now reproduce in other applications what appears to 
be the problem - Two cpu hogs, X and evolution for example are running and 
evolution is making X the cpu hog. The problem is that X gets demoted whereas 
evolution doesn't. Strangely, dropping evolution to nice +1 or making X -1 
seems to change which one gets demoted, and X is now much smoother. I assume 
the same thing is happening here between wine and wineserver, which is why 
you've seen reversal of priorities in your previous posts. See if renicing 
one of them +1 helps for the time being. I will continue investigating to 
find out why the heck this happens and try and fix it.

Con

P.S. I've cc'ed MG because he has seen the scheduler do other forms of 
trickery and may have thoughts on why this happens.