21 Jan
2005
21 Jan
'05
11:32 a.m.
On Fri, 21 Jan 2005 14:52 , Florian Schmidt <mista.tapas(a)gmx.net> sent:
On Fri, 21 Jan 2005 08:12:27 -0600
Jan Depner eviltwin69(a)cableone.net> wrote:
Yes. Take a look at the filesystem tests section on this page. Mark Knecht did
a few tests and the results are very interesting.
http://myweb.cableone.net/eviltwin69/Arcana.html
Jan
On Fri, 2005-01-21 at 07:00, Florian Schmidt
wrote:
what kind of problems? You mean, because the journal is on another part
of the disk, the head has to move a lot while writing? (writing data -
updating journal - writing data - updating journal...) On Thu, 20 Jan 2005 20:02:28 -0800 (PST)
R Parker rtp405(a)yahoo.com> wrote:
The hournal is not with the file. This causes problems when the
file is written to. I'm hesitant about returning to ext2 because
I've had
fsck blow up in my face more than once...guess that's
why I rsync to a big IDE HD. I do expect ext2
performance to be perfectly acceptable.
what's wrong with ext3?
flo
--
Palimm Palimm!
21 Jan
21 Jan
12:13 p.m.
New subject: [linux-audio-user] XFS on FC2; experiences, pointers?
On Fri, 21 Jan 2005 08:50:44 -0700
<eviltwin69(a)cableone.net> wrote:
Yes. Take a look at the filesystem tests section on
this page. Mark Knecht did
a few tests and the results are very interesting.
http://myweb.cableone.net/eviltwin69/Arcana.html
Thanks..
I wonder though: How old are these tests? What kernels were used for it?
Skimming through the page i see a 2.4.19 kernel mentioned. Hmm, it would
be interesting to try to see what results a modern 2.6.x RP kernel and ext3
yield.. I'll try and see if i can get that test of his to run..
Flo
--
Palimm Palimm!
http://affenbande.org/~tapas/
1 p.m.
New subject: [linux-audio-user] XFS on FC2; experiences, pointers?
Florian Schmidt <mista.tapas(a)gmx.net> writes:
On Fri, 21 Jan 2005 08:50:44 -0700
<eviltwin69(a)cableone.net> wrote:
> Yes. Take a look at the filesystem tests section on this page.
> Mark Knecht did a few tests and the results are very interesting.
>
> http://myweb.cableone.net/eviltwin69/Arcana.html
I wonder though: How old are these tests? What kernels
were used for it?
Skimming through the page i see a 2.4.19 kernel mentioned. Hmm, it would
be interesting to try to see what results a modern 2.6.x RP kernel and ext3
yield.. I'll try and see if i can get that test of his to run..
That would be very interesting. Mark did a fine piece of work making
these comparisons. His tests were run on 2.4 at -p256 -n3. Since we
can do much lower latencies now, -p64 -n2 would be interesting. As a
baseline, the vanilla 2.6.10 kernel runs pretty well.
Several kernel developers tell me that recent versions of ext3 are
much better in both disk performance and latency. I'd love to see XFS
and JFS numbers, too.
IIRC, Mark was running Benno Senoner's latency test...
http://www.gardena.net/benno/linux/audio/
--
joq
2:23 p.m.
New subject: [linux-audio-user] XFS on FC2; experiences, pointers?
Jack O'Quin wrote:
can do much lower latencies now, -p64 -n2 would be
interesting. As a
baseline, the vanilla 2.6.10 kernel runs pretty well.
...
IIRC, Mark was running Benno Senoner's latency test...
http://www.gardena.net/benno/linux/audio/
http://www.gardena.net/benno/linux/latencytest-0.42.tar.gz
I guess this is the actual test code. I just tried compiling it
on an AMD64 Gentoo machine with 2.6.10 and got these errors...
gcc -Wall -O2 -DUSE_PENTIUM_TIMER -c rtc_latencytest.c
rtc_latencytest.c:24:41: macro "rdtsc" requires 2 arguments, but only 1 given
rtc_latencytest.c:25: error: syntax error before '{' token
Has anyone else had any success with a similar system setup ?
--markc
7275
days inactive
7275
days old
linux-audio-user@lists.linuxaudio.org
3 comments
4 participants
participants (4)
-
eviltwin69＠cableone.net -
Florian Schmidt -
Jack O'Quin -
Mark Constable