[LAU] reiser4.... has it a place for audio?

Kjetil S. Matheussen k.s.matheussen at notam02.no
Sun Apr 13 13:27:01 EDT 2008


On Sun, 13 Apr 2008, Justin Smith wrote:

> On Sun, Apr 13, 2008 at 9:24 AM, Kjetil S. Matheussen <
> k.s.matheussen at notam02.no> wrote:
> 
> >
> > "Lee Revell":
> > > On Fri, Apr 11, 2008 at 11:43 PM, naysayer <gateswideopen at gmail.com>
> > wrote:
> > > > hi crew...
> > > >
> > > > i have just been looking at the benchmarking stats for reiser4 and it
> > looks
> > > > pretty cool. i was wondering if there is there is any reason to
> > implement
> > > > this kind of file system in multimedia environments. i current have my
> > > > root/system partition as reiserFS and my home partition as ext3. i
> > find this
> > > > to be quite efficient but perhaps if there would be an improvement to
> > > > recording stability and speed, then perhaps it could work well with
> > apps
> > > > like ardour. although, reiser seems to be happiest with small files,
> > reiser4
> > > > claims to be more efficient than ext3.... or that could be just spin.
> > > >
> > >
> > > These days the choice of filesystem should make no significant
> > > difference.  Certainly it won't matter if an -rt kernel is used.
> > >
> >
> > Well, doesn't reiserfs use quite a lot more cpu than ext3? I think
> > that may make a significant difference...
> > _______________________________________________
> > Linux-audio-user mailing list
> > Linux-audio-user at lists.linuxaudio.org
> > http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
> >
> 
> Reiserfs is optimized for smaller files (10 times as fast for files under 1k
> size, compared to ext3), and sluggish on huge files, so it would perform
> worse if you are mainly using large data  (multitracking, wavetable
> synthesis), and slightly better if you are mainly working with small data,
> or data created in ram at runtime (fm, am, algorithmic synthesis, real time
> sampling). For a multitract machine reiser would definitely not help, for a
> machine emulating a moog, or doing experimental realtime synthesis, it may
> be a small improvement.
> 

That's already been said.  What I pointed out was that reiserfs uses more
cpu time than ext3, and for audio use, cpu usage may make a significant 
difference, while throughput, which you are talking about, probably 
don't.




More information about the Linux-audio-user mailing list