18 Apr
2007
18 Apr
'07
7:48 p.m.
I built a computer to run Linux and Ardour. I installed a GeForce
7300 GS video card in it. With help from Paul Davis and others on the
LAD forum, I was able to get jack/ardour/hydrogen/etc working. However,
xruns problems make these programs unusable. I think the video card is
causing the xruns, because they go away after about 15 seconds of no
motion with the mouse or files in the GUI. As soon as I move the mouse
or open a popup menu or anything else, the xruns return and sequire
several seconds to go away.
With this card I am also getting long streams when I move things
around on the Windows desktop. The ghosting also happens in Ubuntu, but
is not as severe.
I have read a little about changing interrupt priority at
http://lowlatency.linuxaudio.org/ , but I could use some guidance and
opinions on what to do.
I plan on installing Ubuntu Studio when it comes out in a couple of
days.
I have a Gigabyte GA-965P-DS3 mobo and a Pentium D-930 processor.
Thanks,
Dave
18 Apr
18 Apr
8 p.m.
On 4/18/07, David Froseth <dfro(a)umich.edu> wrote:
I built a computer to run Linux and Ardour. I
installed a GeForce
7300 GS video card in it. With help from Paul Davis and others on the
LAD forum, I was able to get jack/ardour/hydrogen/etc working. However,
xruns problems make these programs unusable. I think the video card is
causing the xruns, because they go away after about 15 seconds of no
motion with the mouse or files in the GUI. As soon as I move the mouse
or open a popup menu or anything else, the xruns return and sequire
several seconds to go away.
If you're using the proprietary nvidia driver, try removing it.
Lee
10:57 p.m.
nvdia proprietary drivers work perfectly. it is much more likely to be irq
sharing. you are using rt kernel ? what does cat /proc/interupts show ?
g.
19 Apr
19 Apr
4:06 a.m.
On Thursday 19 April 2007, Geoff Beasley wrote:
nvdia proprietary drivers work perfectly.
haha, that's quite a bland statement.. It's always good to try with an
untainted kernel.
it is much more likely to be irq
sharing. you are using rt kernel ? what does cat /proc/interupts show ?
Flo
--
Palimm Palimm!
http://tapas.affenbande.org
4:13 a.m.
On Thursday 19 April 2007 16:06:31 Florian Schmidt wrote:
On Thursday 19 April 2007, Geoff Beasley wrote:
i have never,never,never had any trouble what-so-ever with stable
nvidia
drivers. my daw has a triple-head nvidia and the other has a dual. completely
transparent.
just so there is no grey area here :)
g.
nvdia proprietary drivers work perfectly.
haha, that's quite a bland statement.. It's always good to try with an
untainted kernel. 
7:18 a.m.
On 4/19/07, Geoff Beasley <songshop(a)bizmedia.com.au> wrote:
On Thursday 19 April 2007 16:06:31 Florian Schmidt
wrote:
which is more than i can say about the quite dodgy xorg nv driver.
Fullscreen slows my machine to crawl with the nv driver and works fine
with nvidia one. I'm using nv driver now though, so i don't get picked
on by Lee ;)
Loki
On Thursday 19 April 2007, Geoff Beasley wrote:
i have never,never,never had any trouble what-so-ever with
stable nvidia
drivers. my daw has a triple-head nvidia and the other has a dual.
completely
transparent.
just so there is no grey area here :)
g.
nvdia proprietary drivers work perfectly.
haha, that's quite a bland statement.. It's always good to try with an
untainted kernel. 
noon
Geoff Beasley wrote:
On Thursday 19 April 2007 16:06:31 Florian Schmidt
wrote:
I've never had problems with the nvidia driver either, until I used the
realtime kernel (2.6.20-rt8) and the MX4000. The latest nvidia driver
doesn't support this chipset. There is a patch for the rt kernel and the
nvidia v9631 drivers (the last version to support the MX4000) but I
couldn't get it to work (I didn't have time to try very hard, though.)
--
brad fuller
http://www.Sonaural.com/
+1 (408) 799-6124
On Thursday 19 April 2007, Geoff Beasley wrote:
i have never,never,never had any trouble what-so-ever with stable nvidia
drivers. my daw has a triple-head nvidia and the other has a dual. completely
transparent.
just so there is no grey area here :)
nvdia proprietary drivers work perfectly.
haha, that's quite a bland statement.. It's always good to try with
an
untainted kernel.
12:25 p.m.
On Thursday 19 April 2007, Brad Fuller wrote:
Geoff Beasley wrote:
To compile the driver, one needs to either restore the "transitional" #define
SA_.... stuff to interrupts.h (the rt8 patch had no business removing them!)
or apply a patch to the extracted nvidia sources and then compile that.
What rt8 kernel patch are you referring to and where does one get it?
On Thursday 19 April 2007 16:06:31 Florian
Schmidt wrote:
I've never had problems with the nvidia driver either, until I used the
realtime kernel (2.6.20-rt8) and the MX4000. The latest nvidia driver
doesn't support this chipset. There is a patch for the rt kernel and the
nvidia v9631 drivers (the last version to support the MX4000) but I
couldn't get it to work (I didn't have time to try very hard, though.) On Thursday 19 April 2007, Geoff Beasley wrote:
i have never,never,never had any trouble what-so-ever with stable nvidia
drivers. my daw has a triple-head nvidia and the other has a dual.
completely transparent.
just so there is no grey area here :) nvdia proprietary drivers work perfectly.
haha, that's quite a bland statement.. It's always good to try with an
untainted kernel.
12:30 p.m.
David Baron wrote:
On Thursday 19 April 2007, Brad Fuller wrote:
gentoo pro audio overlay
--
brad fuller
http://www.Sonaural.com/
+1 (408) 799-6124
Geoff Beasley wrote:
To compile the driver, one needs to either restore the "transitional" #define
SA_.... stuff to interrupts.h (the rt8 patch had no business removing them!)
or apply a patch to the extracted nvidia sources and then compile that.
What rt8 kernel patch are you referring to and where does one get it?
On Thursday 19 April 2007 16:06:31 Florian
Schmidt wrote:
I've never had problems with the nvidia driver either, until I used
the
realtime kernel (2.6.20-rt8) and the MX4000. The latest nvidia driver
doesn't support this chipset. There is a patch for the rt kernel and the
nvidia v9631 drivers (the last version to support the MX4000) but I
couldn't get it to work (I didn't have time to try very hard, though.)
On Thursday 19 April 2007, Geoff Beasley wrote:
> nvdia proprietary drivers work perfectly.
>
haha, that's quite a bland statement.. It's always good to try with an
untainted kernel.
i have never,never,never had any trouble what-so-ever with stable nvidia
drivers. my daw has a triple-head nvidia and the other has a dual.
completely transparent.
just so there is no grey area here :)
1:02 p.m.
On Thursday 19 April 2007, Brad Fuller wrote:
gentoo pro audio overlay
Don't help me with Debian :-)
I've never had problems with the nvidia driver
either, until I used the
realtime kernel (2.6.20-rt8) and the MX4000. The latest nvidia driver
doesn't support this chipset. There is a patch for the rt kernel and the
nvidia v9631 drivers (the last version to support the MX4000) but I
couldn't get it to work (I didn't have time to try very hard, though.)
To compile the driver, one needs to either restore the "transitional"
#define SA_.... stuff to interrupts.h (the rt8 patch had no business
removing them!) or apply a patch to the extracted nvidia sources and then
compile that.
What rt8 kernel patch are you referring to and where does one get it?
24 Apr
24 Apr
9:14 a.m.
On Do, 19.04.07 18:02 David Baron <d_baron(a)012.net.il> wrote:
On Thursday 19 April 2007, Brad Fuller wrote:
gentoo pro audio overlay
Don't help me with Debian :-)
http://people.redhat.com/~mingo/realtime-preempt/older/patch-2.6.20-rt8
and the patch for nvidia:
http://gimpel.ath.cx/~tom/files/portage-overlay/gimpel/x11-drivers/nvidia-d…
Cheers.
I've never had problems with the nvidia driver
either, until I
used the realtime kernel (2.6.20-rt8) and the MX4000. The latest
nvidia driver doesn't support this chipset. There is a patch for
the rt kernel and the nvidia v9631 drivers (the last version to
support the MX4000) but I couldn't get it to work (I didn't have
time to try very hard, though.)
To compile the driver, one needs to either restore the
"transitional" #define SA_.... stuff to interrupts.h (the rt8
patch had no business removing them!) or apply a patch to the
extracted nvidia sources and then compile that.
What rt8 kernel patch are you referring to and where does one get
it? 
19 Apr
19 Apr
9:05 a.m.
Geoff Beasley wrote:
nvdia proprietary drivers work perfectly. it is much
more likely to be irq
sharing. you are using rt kernel ? what does cat /proc/interupts show ?
I've used nVidia cards exclusively for many years. I've used the kernel
nv module and nVidia's proprietary binaries. IMO, as long as 3D isn't
required you can use the nv module without much concern. If you need 3D
your only choice is the binary module.
As Geoff mentions, the problems I've had with video cards have had more
to do with IRQ sharing than with the cards or drivers. It's amazing how
badly-built some mobos can be.
Best,
dp
12:16 p.m.
On 4/18/07, Geoff Beasley <songshop(a)bizmedia.com.au> wrote:
nvdia proprietary drivers work perfectly. it is much
more likely to be irq
sharing. you are using rt kernel ? what does cat /proc/interupts show ?
I realize nv driver is often not an option and was just suggesting to
disable proprietary driver for debug purposes, to isolate the cause of
the xruns.
Lee
7:18 p.m.
On Friday 20 April 2007 00:16:47 Lee Revell wrote:
I realize nv driver is often not an option and was
just suggesting to
disable proprietary driver for debug purposes, to isolate the cause of
the xruns.
and i wasn't questioning that ;) just that everytime i've seen this behaviour
it has never had anything to do with the gpu driver layer., tho it's
certainly involved with the problem . changing the gpu driver may just gonna
confirm the obvious.. it's involved but not the cause.
Dave hit the nail.. f*****g mobo's are generally to blame. irq anarchy.
do the slot-shuffle, sit back and watch the show :). you'll probably have to
share your gpu irq, but do it with a nic...your audio card should not be
sharing for best performance. use xt-pic if you can.
g.
20 Apr
20 Apr
4:28 a.m.
Thanks for all of the feedback.
The contents of /proc/interrupts is:
~$ cat /proc/interrupts
CPU0 CPU1
0: 668825 0 IO-APIC-edge timer
1: 7 0 IO-APIC-edge i8042
7: 0 0 IO-APIC-edge parport0
8: 3 0 IO-APIC-edge rtc
9: 1 0 IO-APIC-level acpi
50: 0 0 IO-APIC-level uhci_hcd:usb5, ehci_hcd:usb6
58: 2 0 IO-APIC-level ohci1394
66: 11431 0 PCI-MSI sky2
74: 782 0 IO-APIC-level HDA Intel
169: 7970 0 IO-APIC-level ide0, uhci_hcd:usb1
177: 69956 0 IO-APIC-level libata, libata, libata,
uhci_hcd:usb4
225: 0 0 IO-APIC-level uhci_hcd:usb2
233: 2 0 IO-APIC-level uhci_hcd:usb3, ehci_hcd:usb7
NMI: 0 0
LOC: 668640 668640
ERR: 0
MIS: 0
I do not undersand how to read all of the info in this printout. Can
anybody tell me where the video card and the mobo's sound card are in
this list? I have ordered a M-Audio Delta 1010 sound card, which is on
the way. But currently I am using the mobo's built in sound card.
Most of the feedback from all of you is above my head at this point.
Can anybody guide me through getting the video behind the sound card in
the IRQ? Should I wait until I get the Delta 1010 before I start all of
this?
I set up the kernel using this ubuntu guide:
https://help.ubuntu.com/community/UbuntuStudioPreparation
Thanks
Dave
Geoff Beasley wrote:
nvdia proprietary drivers work perfectly. it is much
more likely to be irq
sharing. you are using rt kernel ? what does cat /proc/interupts show ?
g.
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-user
5:07 a.m.
On Friday 20 April 2007 16:28:00 David Froseth wrote:
ok, 8 usb's, firewire, sata, etc and dual core or dual cpu. your system has a
lot of devices but not one of them is your gpu. if you're using the nv driver
then that would explain it as it doesn't use an intetrupt ( if i remember
correctly. correct me if i'm wrong ;) )
I do not undersand how to read all of the info in this
printout. Can
anybody tell me where the video card and the mobo's sound card are in
your soundcard is the hda device. if you're running jack with a period size of
2 that may account for your xruns behaviour. these chips are notorious.
...change it to 3.
also ensure your you are running jack with a priority of around 79. i assume
you use qjackctl ? both of these parameters are on the opening page under
setup>settings.... also use 48khz sample rate.. otherwise it may be
re-sampling on the fly.
this list? I have ordered a M-Audio Delta 1010 sound
card, which is on
the way. But currently I am using the mobo's built in sound card.
good choice. very well supported :)
Most of the feedback from all of you is above my head
at this point.
Can anybody guide me through getting the video behind the sound card in
the IRQ? Should I wait until I get the Delta 1010 before I start all of
this?
you have to use apic, so interupt sharing is just gonna happen. apic is
getting better tho, so with just one card you may find it will work fine....
and get that nvidia driver back in there when you get your 1010.
hth,
g.
6232
days inactive
6238
days old
linux-audio-user@lists.linuxaudio.org
16 comments
9 participants
participants (9)
-
Brad Fuller -
Dave Phillips -
David Baron -
David Froseth -
Florian Schmidt -
Geoff Beasley -
Lee Revell -
Loki Davison -
Thomas Kuther