31 Jan
2007
31 Jan
'07
6:25 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
I've run into an 80% DSP problem on my machine.
This poor little PC doesn't have enough CPU power to run all the softsynths and LADSPA
plugins I tend to use. When JACK gets to 80% DSP usage, all hell breaks loose. If I up the
latency, then my CPU usage goes down, but the system is unplayable (too much latency).
So. What kinds of things would be best to try to squeeze some life out of this
underpowered 1.66Ghz Core Duo? I have a list, but I'm not sure which would be the
first, second, third order improvements and I'd like to try to conserve time,
especially if the end result will end up being that I have to sell this PC and get a new
one anyway. Do I have the priorities right?
1) Try jackdmp instead of jackd
2) Try DRM or some kind of accelerated graphics for Xorg
3) Blindly chase "latest and greatest" versions of things like kernel 2.4.20,
latest jackd, svn freebob, etc.
4) Try messing around with PCI bus latency
I'll try as many of those as you-all think might be worth pursuing, before moving on
to hardware solutions:
5) Replace the CPU on this box with a 2.16Ghz Core Duo.
6) Get rid of the machine completely and get a new PC with 2Ghz+ Intel or AMD dual CPU,
and PCI audio card not firewire.
- -ken
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFFwG6We8HF+6xeOIcRApf6AKCmlNbp8oWhGP2THi+va2VqfKWGiACgqitk
NWSBmNwyws/atURFGYO+FyE=
=MouT
-----END PGP SIGNATURE-----
31 Jan
31 Jan
11:52 p.m.
On Wed, Jan 31, 2007 at 02:25:26AM -0800, Ken Restivo wrote:
2) Try DRM or some kind of accelerated graphics for
Xorg
This might make the problem worse. Many graphics cards are mean to the
PCI bus. The extra interrupts can make latency issues worse.
What version of fftw is being used on your system? If you're using
fftw v2, make sure to build it for your arch! If you're using fftw3,
arch specific SIMD is automatically used by the lib.
--
Ross Vandegrift
ross(a)kallisti.us
"The good Christian should beware of mathematicians, and all those who
make empty prophecies. The danger already exists that the mathematicians
have made a covenant with the devil to darken the spirit and to confine
man in the bonds of Hell."
--St. Augustine, De Genesi ad Litteram, Book II, xviii, 37
1 Feb
1 Feb
7:07 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wed, Jan 31, 2007 at 10:52:50PM -0500, Ross Vandegrift wrote:
On Wed, Jan 31, 2007 at 02:25:26AM -0800, Ken Restivo
wrote:
This is getting really frustrating. I'm getting shunted from pillar to post by
conflicting advice from all sides. I'm lost in the ambiguity. This is a "Intel
Corporation Mobile 945GM/GMS/940GML Express Integrated Graphics Controller". How
would I find out if it is one of the "many"?
2) Try DRM or some kind of accelerated graphics
for Xorg
This might make the problem worse. Many graphics cards are mean to the
PCI bus. The extra interrupts can make latency issues worse.
What version of fftw is being used on your system? If you're using
fftw v2, make sure to build it for your arch! If you're using fftw3,
arch specific SIMD is automatically used by the lib.
$ dpkg -l |grep fftw
ii fftw2 2.1.3-20 library for
computing Fast Fourier Transforms
ii fftw3 3.1.2-1 library for
computing Fast Fourier Transforms
ii sfftw2 2.1.3-20 library for
computing Fast Fourier Transforms
It looks like the only package I have installed that uses it is rezound, which is indeed
very slow on my system, but isn't typically in my signal chain.
Interestingly, AMS, which *is* slow and *is* in my signal chain, uses sfftw2 (not sure
what's the difference between fftw2 and sfftw2), so I'll definitely try rebuilding
it and optimizing it for this CPU (Pentium M, according to kernel config, and vendor_id:
GenuineIntel, cpu family: 6, model: 14, according to /proc/cpuinfo).
- -ken
- --
P.S: a handy one-liner for Debian to find which installed packages use fftw2:
(for i in `apt-cache showpkg fftw2 | grep ','| cut -d, -f 1`; do dpkg -s $i; done
2>/dev/null) |grep Package:
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFFwnKNe8HF+6xeOIcRApMDAJ9mhbWZku4HWPMagwgnHLj3HnBlhQCfaqFA
21iEvhmTA3p5PfREw7VREYw=
=SB/D
-----END PGP SIGNATURE-----
3 Feb
3 Feb
12:32 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wed, Jan 31, 2007 at 10:52:50PM -0500, Ross Vandegrift wrote:
On Wed, Jan 31, 2007 at 02:25:26AM -0800, Ken Restivo
wrote:
Speaking of making the problem worse, using the p4fftwgel2 package, which is optimized for
P4, makes CPU usage *higher*. I'm not even kidding.
But, and even more strangely, it seems to also makes the crackling a little better, even
though CPU usage is higher. It's as if it allows the CPU usage to go to higher levels
before failure, but itself takes up more CPU.
I'm completely baffled as to why this is.
- -ken
2) Try DRM or some kind of accelerated graphics
for Xorg
This might make the problem worse. Many graphics cards are mean to the
PCI bus. The extra interrupts can make latency issues worse.
What version of fftw is being used on your system? If you're using
fftw v2, make sure to build it for your arch! If you're using fftw3,
arch specific SIMD is automatically used by the lib.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFFxBBAe8HF+6xeOIcRApJZAKCh5Q7IFZZQvxyCFd+82m8o3iJuGQCfbc2a
XWM1nLIKB1yhGncNyiBmJMg=
=zP13
-----END PGP SIGNATURE-----
6293
days inactive
6295
days old
linux-audio-user@lists.linuxaudio.org
3 comments
2 participants
participants (2)
-
Ken Restivo -
Ross Vandegrift