25 Nov
2004
25 Nov
'04
3:17 a.m.
Something apparently unrelated to Linux audio, but in fact it is. Let me
explain.
1. Is there a way to run an NVidia card dual-head (two monitors) using
the open source drivers?
2. Is there a way to silence the fan on an NVidia-based card other than
loading up the proprietary drivers? (which silences it as long as there
is no hard graphics work)
Being able to extend the desktop across two monitors, and not having to
bear the obnoxious fan on the graphics card are the reasons why i'm
using the NVidia proprietary driver, because:
- i want to use two monitors because audio apps are hungry for desktop
real estate
- the noisy fan is very disturbing when doing any kind of audio work
But there's a nasty side-effect of the proprietary driver - the system
seems to freeze for a few seconds every once in a while (about once a
day).
There are no side-effects of the open source driver, but i don't know
how to tell it to work in dual-head mode and i don't know how to silence
the fan.
--
Florin Andrei
http://florin.myip.org/
25 Nov
25 Nov
4:16 a.m.
We've been using the NVIDIA cards and drivers at work on our smp and
non-smp systems (all dual monitor) for a couple of years. There is
still an insidious little bug somewhere in the driver that only shows up
on smp systems that causes the behavior that you are experiencing. We
only see it about once a month though and our systems are up 24/7. We
don't have any problems with non-smp systems. From what I've read
running dual monitors without the NVIDIA driver is a massive PITA if
it's even possible. I think you have to do some fancy foot work with
Xinerama to get it to work.
Jan
On Thu, 2004-11-25 at 01:17, Florin Andrei wrote:
Something apparently unrelated to Linux audio, but in
fact it is. Let me
explain.
1. Is there a way to run an NVidia card dual-head (two monitors) using
the open source drivers?
2. Is there a way to silence the fan on an NVidia-based card other than
loading up the proprietary drivers? (which silences it as long as there
is no hard graphics work)
Being able to extend the desktop across two monitors, and not having to
bear the obnoxious fan on the graphics card are the reasons why i'm
using the NVidia proprietary driver, because:
- i want to use two monitors because audio apps are hungry for desktop
real estate
- the noisy fan is very disturbing when doing any kind of audio work
But there's a nasty side-effect of the proprietary driver - the system
seems to freeze for a few seconds every once in a while (about once a
day).
There are no side-effects of the open source driver, but i don't know
how to tell it to work in dual-head mode and i don't know how to silence
the fan.
6:43 a.m.
Hi,
Now that we are talking nvidia, I take advantage to ask if anyone got
the nvidia proprietary modules to build correctly with an "Ingo Molnars
realtime preemption patch" patched kernel?
maarten
7:12 a.m.
Maarten de Boer wrote:
Hi,
Now that we are talking nvidia, I take advantage to ask if anyone got
the nvidia proprietary modules to build correctly with an "Ingo Molnars
realtime preemption patch" patched kernel?
maarten
Yes, while the patches were based against 2.6.9. Since 2.6.10-rc1 the
proprietary nvidia kernel modules refuse to load due to kernel symbols
mismatch. At least on me.
AFAICT it's not Ingo's RT patch fault, it's a Linus/Andrew's feature :)
Either way, I guess nvidia should do it's homework, once 2.6.10 is out.
Bye.
--
rncbc aka Rui Nuno Capela
rncbc(a)rncbc.org
7:18 a.m.
Maarten de Boer skrev:
Hi,
Now that we are talking nvidia, I take advantage to ask if anyone got
the nvidia proprietary modules to build correctly with an "Ingo Molnars
realtime preemption patch" patched kernel?
maarten
yes i have
The srpm is available here http://rpm.nyvalls.se/graphics10.1.html
/Torbjörn Turpeinen (Thac)
7:41 a.m.
On Thu, 25 Nov 2004 12:18:34 +0100
"tobbe(a)nyvalls.se" <tobbe(a)nyvalls.se> wrote:
Maarten de Boer skrev:
The nvidia binary only driver has some nasty latencies (around 400us),
so i wouldn't recommend using it for low latency work anyways. I don't
know if the realtime preemptive kernel can handle it better than the
older voluntary preemption kernels due to their design. It would be
interesting to get some data on that.
Would you mind providing a patched tarball of the nvidia driver (since i
don't use an rpm based distro).
flo
Hi,
Now that we are talking nvidia, I take advantage to ask if anyone got
the nvidia proprietary modules to build correctly with an "Ingo Molnars
realtime preemption patch" patched kernel?
maarten
yes i have
The srpm is available here http://rpm.nyvalls.se/graphics10.1.html
3:52 p.m.
On Thu, 2004-11-25 at 02:35 -0600, Jan Depner wrote:
We've been using the NVIDIA cards and drivers at
work on our smp and
non-smp systems (all dual monitor) for a couple of years. There is
still an insidious little bug somewhere in the driver that only shows up
on smp systems that causes the behavior that you are experiencing. We
only see it about once a month though and our systems are up 24/7. We
don't have any problems with non-smp systems.
Actually, my system is a single-CPU AthlonXP 2600+ running Fedora Core 3
(distribution-provided kernel and all).
--
Florin Andrei
http://florin.myip.org/
4:19 p.m.
On Thu, 2004-11-25 at 13:52, Florin Andrei wrote:
On Thu, 2004-11-25 at 02:35 -0600, Jan Depner wrote:
Interesting. My home system is an Athlon 1700+ running FC1/Planet
kernel with NVIDIA card and NVIDIA driver. It is also up 24/7 and I
never have a single problem with the graphics/video. When running
Ardour I do kick the NVIDIA PCI latency timer down to 32 instead of the
default (from the driver) 248.
Jan
We've been using the NVIDIA cards and drivers
at work on our smp and
non-smp systems (all dual monitor) for a couple of years. There is
still an insidious little bug somewhere in the driver that only shows up
on smp systems that causes the behavior that you are experiencing. We
only see it about once a month though and our systems are up 24/7. We
don't have any problems with non-smp systems.
Actually, my system is a single-CPU AthlonXP 2600+ running Fedora Core 3
(distribution-provided kernel and all). 
1:19 p.m.
My audio system is running 2 fanless ATI cards, one AGP and one PCI. I
would be happy to send my xfree config files off list if you like. I am
using fedora core 1 which set it all up automagically for me. I had
heaps of trouble using an older distro (hence an older xfree) to get
dual head behaviour from my ATI cards and from an NVidia card.
A fanless card goes a long way! There is a water cooled system made by
Zalman which allows you to replace the video fans with water cooled
pipes.
-Joe D
On Thu, 2004-11-25 at 02:17, Florin Andrei wrote:
Something apparently unrelated to Linux audio, but in
fact it is. Let me
explain.
1. Is there a way to run an NVidia card dual-head (two monitors) using
the open source drivers?
2. Is there a way to silence the fan on an NVidia-based card other than
loading up the proprietary drivers? (which silences it as long as there
is no hard graphics work)
Being able to extend the desktop across two monitors, and not having to
bear the obnoxious fan on the graphics card are the reasons why i'm
using the NVidia proprietary driver, because:
- i want to use two monitors because audio apps are hungry for desktop
real estate
- the noisy fan is very disturbing when doing any kind of audio work
But there's a nasty side-effect of the proprietary driver - the system
seems to freeze for a few seconds every once in a while (about once a
day).
There are no side-effects of the open source driver, but i don't know
how to tell it to work in dual-head mode and i don't know how to silence
the fan.
7331
days inactive
7331
days old
linux-audio-user@lists.linuxaudio.org
8 comments
7 participants
participants (7)
-
Florian Schmidt -
Florin Andrei -
Jan Depner -
Joseph Dell'Orfano -
Maarten de Boer -
Rui Nuno Capela -
tobbe＠nyvalls.se