Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Andrew Morton]

On Sat, 19 Aug 2006 11:00:22 -0700
bugme-daemon@bugzilla.kernel.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=7027
> 
>            Summary: CD Ripping speeds slow with 2.6.17
>     Kernel Version: 2.6.17
>             Status: NEW
>           Severity: normal
>              Owner: bzolnier@gmail.com
>          Submitter: brnewber@gmail.com
> 
> 
> Most recent kernel where this bug did not occur: 2.6.16
> Distribution: Gentoo
> Hardware Environment: ASUS K8V mobo, AMD64 2200, 1GB RAM
> Software Environment: Gentoo
> Problem Description: Ever since 2.6.17 my cd ripping speeds with one particular
> CD ripper/encoder (namely the one I wrote) have been slow. 
> 
> Using git bisect I tracked it down to this patch..
> 
> 9430d58e34ec3861e1ca72f8e49105b227aad327 is first bad commit
> commit 9430d58e34ec3861e1ca72f8e49105b227aad327
> Author: Mike Galbraith <efault@gmx.de>
> Date:   Wed Mar 22 00:07:33 2006 -0800
> 
>     [PATCH] sched: remove sleep_avg multiplier
> 
>     Remove the sleep_avg multiplier.  This multiplier was necessary back when
>     we had 10 seconds of dynamic range in sleep_avg, but now that we only have
>     one second, it causes that one second to be compressed down to 100ms in
>     some cases.  This is particularly noticeable when compiling a kernel in a
>     slow NFS mount, and I believe it to be a very likely candidate for other
>     recently reported network related interactivity problems.
> 
>     In testing, I can detect no negative impact of this removal.
> 
>     Signed-off-by: Mike Galbraith <efault@gmx.de>
>     Acked-by: Ingo Molnar <mingo@elte.hu>
>     Signed-off-by: Andrew Morton <akpm@osdl.org>
>     Signed-off-by: Linus Torvalds <torvalds@osdl.org>
> 
> :040000 040000 28d2d8f53ab7b5dd89e846f2dcc107ce88cb695f 780a13c0f8ba5465db79c668
> 
> I'm honestly not sure if my application is doing something it shouldn't or if
> this is a legitimate kernel bug. Being totally at a loss I'm filing it here.I
> just know that before the above patch I was ripping at about speeds of 9.0x and
> now I rip at 1.2x.
> 

sched problems...

Martin, this might be the source of your `dvdrip' woes?

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[mbligh]

Andrew Morton wrote:
> On Sat, 19 Aug 2006 11:00:22 -0700
> bugme-daemon@bugzilla.kernel.org wrote:
> 
>> http://bugzilla.kernel.org/show_bug.cgi?id=7027
>>
>>            Summary: CD Ripping speeds slow with 2.6.17
>>     Kernel Version: 2.6.17
>>             Status: NEW
>>           Severity: normal
>>              Owner: bzolnier@gmail.com
>>          Submitter: brnewber@gmail.com
>>
>>
>> Most recent kernel where this bug did not occur: 2.6.16
>> Distribution: Gentoo
>> Hardware Environment: ASUS K8V mobo, AMD64 2200, 1GB RAM
>> Software Environment: Gentoo
>> Problem Description: Ever since 2.6.17 my cd ripping speeds with one particular
>> CD ripper/encoder (namely the one I wrote) have been slow. 
>>
>> Using git bisect I tracked it down to this patch..
>>
>> 9430d58e34ec3861e1ca72f8e49105b227aad327 is first bad commit
>> commit 9430d58e34ec3861e1ca72f8e49105b227aad327
>> Author: Mike Galbraith <efault@gmx.de>
>> Date:   Wed Mar 22 00:07:33 2006 -0800
>>
>>     [PATCH] sched: remove sleep_avg multiplier
>>
>>     Remove the sleep_avg multiplier.  This multiplier was necessary back when
>>     we had 10 seconds of dynamic range in sleep_avg, but now that we only have
>>     one second, it causes that one second to be compressed down to 100ms in
>>     some cases.  This is particularly noticeable when compiling a kernel in a
>>     slow NFS mount, and I believe it to be a very likely candidate for other
>>     recently reported network related interactivity problems.
>>
>>     In testing, I can detect no negative impact of this removal.
>>
>>     Signed-off-by: Mike Galbraith <efault@gmx.de>
>>     Acked-by: Ingo Molnar <mingo@elte.hu>
>>     Signed-off-by: Andrew Morton <akpm@osdl.org>
>>     Signed-off-by: Linus Torvalds <torvalds@osdl.org>
>>
>> :040000 040000 28d2d8f53ab7b5dd89e846f2dcc107ce88cb695f 780a13c0f8ba5465db79c668
>>
>> I'm honestly not sure if my application is doing something it shouldn't or if
>> this is a legitimate kernel bug. Being totally at a loss I'm filing it here.I
>> just know that before the above patch I was ripping at about speeds of 9.0x and
>> now I rip at 1.2x.
>>
> 
> sched problems...
> 
> Martin, this might be the source of your `dvdrip' woes?
> 

Nope, was an older kernel than that, and oddly doesn't always seem to 
happen. I'll try to get that box upgraded, and try to reproduce it
again.

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Mike Galbraith]

On Sat, 2006-08-19 at 11:14 -0700, Andrew Morton wrote:
> On Sat, 19 Aug 2006 11:00:22 -0700
> bugme-daemon@bugzilla.kernel.org wrote:
> 
> > http://bugzilla.kernel.org/show_bug.cgi?id=7027
> > 
> >            Summary: CD Ripping speeds slow with 2.6.17
> >     Kernel Version: 2.6.17
> >             Status: NEW
> >           Severity: normal
> >              Owner: bzolnier@gmail.com
> >          Submitter: brnewber@gmail.com
> > 
> > 
> > Most recent kernel where this bug did not occur: 2.6.16
> > Distribution: Gentoo
> > Hardware Environment: ASUS K8V mobo, AMD64 2200, 1GB RAM
> > Software Environment: Gentoo
> > Problem Description: Ever since 2.6.17 my cd ripping speeds with one particular
> > CD ripper/encoder (namely the one I wrote) have been slow. 
> > 
> > Using git bisect I tracked it down to this patch..
> > 
> > 9430d58e34ec3861e1ca72f8e49105b227aad327 is first bad commit
> > commit 9430d58e34ec3861e1ca72f8e49105b227aad327
> > Author: Mike Galbraith <efault@gmx.de>
> > Date:   Wed Mar 22 00:07:33 2006 -0800
> > 
> >     [PATCH] sched: remove sleep_avg multiplier
> > 
> >     Remove the sleep_avg multiplier.  This multiplier was necessary back when
> >     we had 10 seconds of dynamic range in sleep_avg, but now that we only have
> >     one second, it causes that one second to be compressed down to 100ms in
> >     some cases.  This is particularly noticeable when compiling a kernel in a
> >     slow NFS mount, and I believe it to be a very likely candidate for other
> >     recently reported network related interactivity problems.
> > 
> >     In testing, I can detect no negative impact of this removal.
> > 
> >     Signed-off-by: Mike Galbraith <efault@gmx.de>
> >     Acked-by: Ingo Molnar <mingo@elte.hu>
> >     Signed-off-by: Andrew Morton <akpm@osdl.org>
> >     Signed-off-by: Linus Torvalds <torvalds@osdl.org>
> > 
> > :040000 040000 28d2d8f53ab7b5dd89e846f2dcc107ce88cb695f 780a13c0f8ba5465db79c668
> > 
> > I'm honestly not sure if my application is doing something it shouldn't or if
> > this is a legitimate kernel bug. Being totally at a loss I'm filing it here.I
> > just know that before the above patch I was ripping at about speeds of 9.0x and
> > now I rip at 1.2x.

It's pretty difficult imagining this patch being responsible for an IO
regression.  The scheduling advantages for applications which sleep even
a little is still absolutely massive.
  
> sched problems...

I'm skeptical.  Is the source for this application available?  I'd like
to see this problem.

	-Mike

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Mike Galbraith]

On Sun, 2006-08-20 at 08:27 +0000, Mike Galbraith wrote:

> I'm skeptical.  Is the source for this application available?  I'd like
> to see this problem.

(never mind.  saw your other post, found source)

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Mike Galbraith]

On Sun, 2006-08-20 at 10:03 +0000, Mike Galbraith wrote:
> On Sun, 2006-08-20 at 08:27 +0000, Mike Galbraith wrote:
> 
> > I'm skeptical.  Is the source for this application available?  I'd like
> > to see this problem.
> 
> (never mind.  saw your other post, found source)

Hm.  I can't get better than 1.4x rip speed out of it with a stock SuSE
10.1 kernel (2.6.16).  It's also using truckloads of cpu, whereas the CD
rippers that came with this distro use a percent or two.

	-Mike

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Jan Engelhardt]

>> > I'm skeptical.  Is the source for this application available?  I'd like
>> > to see this problem.
>> 
>> (never mind.  saw your other post, found source)
>
>Hm.  I can't get better than 1.4x rip speed out of it with a stock SuSE
>10.1 kernel (2.6.16).  It's also using truckloads of cpu, whereas the CD
>rippers that came with this distro use a percent or two.

What command did you use to rip?



Jan Engelhardt
-- 

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Ryan Newberry]

Jan Engelhardt wrote:
>>>> I'm skeptical.  Is the source for this application available?  I'd like
>>>> to see this problem.
>>>>         
>>> (never mind.  saw your other post, found source)
>>>       
>> Hm.  I can't get better than 1.4x rip speed out of it with a stock SuSE
>> 10.1 kernel (2.6.16).  It's also using truckloads of cpu, whereas the CD
>> rippers that came with this distro use a percent or two.
>>     
>
> What command did you use to rip?
>
>
>
> Jan Engelhardt
>   
The ripper he's using is ripoff I assume (source code here: 
http://prdownloads.sourceforge.net/ripoffc/ripoff-0.8.tar.gz?download  
extraction functionality contained in src/RipOffExtractor.c) . It uses 
libparanoia to do its job, like the cdparanoia command. On my system, 
ripoff has high CPU usage with a 2.6.16 kernel as well, but it reports a 
9.0x rate on average.
Could the fact that it has such high CPU usage be a possible reason I am 
experiencing a slower ripping speed (1.2x) when the patch that was git 
bisected is applied?

-- 
Ryan Newberry
http://ripoffc.sourceforge.net
"All mankind is divided into three classes: those that are immovable, those that are movable, and those that move." - Benjamin Franklin

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Mike Galbraith]

On Sun, 2006-08-20 at 13:28 -0400, Ryan Newberry wrote:
> Jan Engelhardt wrote:
> >>>> I'm skeptical.  Is the source for this application available?  I'd like
> >>>> to see this problem.
> >>>>         
> >>> (never mind.  saw your other post, found source)
> >>>       
> >> Hm.  I can't get better than 1.4x rip speed out of it with a stock SuSE
> >> 10.1 kernel (2.6.16).  It's also using truckloads of cpu, whereas the CD
> >> rippers that came with this distro use a percent or two.
> >>     
> >
> > What command did you use to rip?
> >
> >
> >
> > Jan Engelhardt
> >   
> The ripper he's using is ripoff I assume (source code here: 
> http://prdownloads.sourceforge.net/ripoffc/ripoff-0.8.tar.gz?download  
> extraction functionality contained in src/RipOffExtractor.c) . It uses 
> libparanoia to do its job, like the cdparanoia command. On my system, 
> ripoff has high CPU usage with a 2.6.16 kernel as well, but it reports a 
> 9.0x rate on average.

Watching the .16 kernel, it boosts to interactive status frequently, but
even with this mega-boost, it uses so much cpu that it repeatedly falls
into the expire (unbounded latency) category.

At this point, I think this app has problems.

> Could the fact that it has such high CPU usage be a possible reason I am 
> experiencing a slower ripping speed (1.2x) when the patch that was git 
> bisected is applied?

Short answer, with a lot of interpolation, yes.

	-Mike

Re: [Bugme-new] [Bug 7027] New: CD Ripping speeds slow with 2.6.17

[Jan Engelhardt]

>> > rippers that came with this distro use a percent or two.
>> 
>> What command did you use to rip?
>> 
> The ripper he's using is ripoff I assume (source code here:

I would suggest trying another one, possibly not based on cdparanoia.


Jan Engelhardt
--