Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 01-13-08, 11:08 AM   #13
mooninite
Registered User
 
Join Date: May 2006
Posts: 477
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by SilentLexx
Don't buy cards of 7 series if you want running Linux on Dual Core/SMP system!
It seems to only affect you and a few other users. I have a 7300GS that works just fine. SMP enabled. Core 2 Duo.

Maybe it's just your 7300LE?
mooninite is offline   Reply With Quote
Old 01-13-08, 11:24 AM   #14
SilentLexx
Registered User
 
SilentLexx's Avatar
 
Join Date: Dec 2007
Location: Ukraine / Sumy
Posts: 112
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

I known guys, who have problems with stability on 7300GS too...
__________________
PS: I'm sorry, but my english is very bad
openSUSE 11.1, Gigabyte GA-M57SLI-S4, AMD Athlon64 X2 +4800@2500MHz, 2 GB RAM, GeForce 9800GT.
openSUSE 11.0, MSI M670X, AMD Sempron +3600@2000MHz, 2 GB RAM, GeForce 6100 Go.
Ubuntu 8.10, Albatron KM266Pro, AMD AthlonXP +1800@1600MHz, 1 GB RAM, GeForce 4 MX440.
SilentLexx is offline   Reply With Quote
Old 01-13-08, 05:12 PM   #15
pferraro
Registered User
 
Join Date: Aug 2007
Posts: 30
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by mooninite
This "bug" doesn't affect all of the 7 series. The 7800 and 7950 cards work fine with composite enabled. 2D, 3D, Xv. Core 2 Duos. SMP enabled.

Just an FYI.
I cannot use compiz for more than a few hours on my GeForce Go 7950 GTX (on 64-bit smp kernel) without X freezing. The NVreg_RegistryDwords="PerfLevelSrc=0x2222" workaround prevents the freezing, but this is not acceptable for a notebook that is prone to overheating. Nor is it acceptable that I run with only 1 core enabled.

Please do not speak for others... It's hard enough getting anyone at nvidia to acknowledge this issue (all of my bug reports have so far been ignored) without someone implying that the problem does not exist.
pferraro is offline   Reply With Quote
Old 01-13-08, 06:37 PM   #16
zarfmouse
Registered User
 
Join Date: Jan 2007
Posts: 11
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

I'm having a problem with the same symptoms on a 8500 GT chipset. I'm using maxcpus=1, idle=poll, and pci=nommconf and none of these solve the problem of random hangs or X-server crashes every 1-4 hours of active use.

My several bug reports and my post (with ongoing updates of my lack of progress) to this forum have not received any response.

I actually would happily cripple my system performance if it meant a way to stabilize my video card but even limiting to one CPU has not helped.
zarfmouse is offline   Reply With Quote
Old 01-13-08, 09:40 PM   #17
mooninite
Registered User
 
Join Date: May 2006
Posts: 477
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by pferraro
I cannot use compiz for more than a few hours on my GeForce Go 7950 GTX (on 64-bit smp kernel) without X freezing. The NVreg_RegistryDwords="PerfLevelSrc=0x2222" workaround prevents the freezing, but this is not acceptable for a notebook that is prone to overheating. Nor is it acceptable that I run with only 1 core enabled.

Please do not speak for others... It's hard enough getting anyone at nvidia to acknowledge this issue (all of my bug reports have so far been ignored) without someone implying that the problem does not exist.
Did you see me say the bug doesn't exist? Please quote me correctly. I said it doesn't affect everyone.

I have 32-bit and 64-bit systems. SMP enabled. 7300GS PCI-E, 7800GS AGP, and 7950GT PCI-E all working with Compiz, 3D, 2D, Xv, you name it. These systems stay up 24/7.

You need to include more details. It's obviously something motherboard (possibly distro) specific.

Always include your bug report... The nVidia group can't do much with out a bug report.
mooninite is offline   Reply With Quote
Old 01-14-08, 12:03 AM   #18
pferraro
Registered User
 
Join Date: Aug 2007
Posts: 30
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by mooninite
Did you see me say the bug doesn't exist?
Yes, actually...
Quote:
The 7800 and 7950 cards work fine with composite enabled.
To me, that statement implies that there is no such bug with the 7950 series (it would not work "fine" if such a bug existed, right?).

Quote:
Originally Posted by mooninite
Please quote me correctly. I said it doesn't affect everyone.
My response quoted your whole message verbatim. So, I couldn't possibly have quoted you incorrectly. I may have *interpretted* your statement incorrectly, and I'm sorry if that was the case. I'm just a little frustrated with how long this bug has been around given how expensive this hardware was...

Quote:
Originally Posted by mooninite
I have 32-bit and 64-bit systems. SMP enabled. 7300GS PCI-E, 7800GS AGP, and 7950GT PCI-E all working with Compiz, 3D, 2D, Xv, you name it. These systems stay up 24/7.
Do you have any extra kernel/driver options tweaked on your 7950 system? As I said before, I can disable GPU frequency scaling via the PerfLevelSrc driver property and gain stability that way, but that's not an unacceptable solution for a notebook with battery and heat concerns. I don't know what the differences are between your 7950 GT and my GTX.

Quote:
Originally Posted by mooninite
You need to include more details. It's obviously something motherboard (possibly distro) specific.
I doubt the issue is distro specific. I've seen many linux users complain about this same issue using Ubuntu, Gentoo, Fedora, you name it. The issue may likely be BIOS related. I'm running the latest version from my manufacturer (Sager NP5760), so I don't know what else to do. So long as nvidia's driver is closed source, it is their responsibility to facilitate bug resolution with motherboard manufacturers.

Quote:
Originally Posted by mooninite
Always include your bug report... The nVidia group can't do much with out a bug report.
I've already sent several bug reports to nvidia regarding this problem. I guess one more can't hurt. Lemme turn compiz on...
pferraro is offline   Reply With Quote
Old 01-14-08, 12:31 AM   #19
SilentLexx
Registered User
 
SilentLexx's Avatar
 
Join Date: Dec 2007
Location: Ukraine / Sumy
Posts: 112
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Btw I downgrade (!) mb BIOS and look like system work more stable. Was other problems with updated BIOS, like HDD not detected on startup. After downgrade all works fine.
But why Windows drivers on my hardware don't have this issue?
I like running Linux, because it's match more stable (and faster!) system then Windows. But NVIDIA drivers do Linux more likly on Windows. Shame!
__________________
PS: I'm sorry, but my english is very bad
openSUSE 11.1, Gigabyte GA-M57SLI-S4, AMD Athlon64 X2 +4800@2500MHz, 2 GB RAM, GeForce 9800GT.
openSUSE 11.0, MSI M670X, AMD Sempron +3600@2000MHz, 2 GB RAM, GeForce 6100 Go.
Ubuntu 8.10, Albatron KM266Pro, AMD AthlonXP +1800@1600MHz, 1 GB RAM, GeForce 4 MX440.
SilentLexx is offline   Reply With Quote
Old 01-14-08, 02:09 PM   #20
mooninite
Registered User
 
Join Date: May 2006
Posts: 477
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by pferraro
Do you have any extra kernel/driver options tweaked on your 7950 system? As I said before, I can disable GPU frequency scaling via the PerfLevelSrc driver property and gain stability that way, but that's not an unacceptable solution for a notebook with battery and heat concerns. I don't know what the differences are between your 7950 GT and my GTX.
No. The machines all run Fedora 8. Fedora stock kernel. No overclocking done. Stock video card BIOS'. Stock clock speeds. My systems do 2D/3D and Compiz with the frequency scaling enabled and never hiccup.

There is no 7950 GTX, however, there is a 7900 GTX. Are you sure what model of card you have?

http://www.nvidia.com/page/geforce_7950.html
http://en.wikipedia.org/wiki/Geforce...ce_7950_Series
mooninite is offline   Reply With Quote

Old 01-14-08, 02:54 PM   #21
pferraro
Registered User
 
Join Date: Aug 2007
Posts: 30
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by mooninite
No. The machines all run Fedora 8. Fedora stock kernel. No overclocking done. Stock video card BIOS'. Stock clock speeds. My systems do 2D/3D and Compiz with the frequency scaling enabled and never hiccup.
Lucky you...

Quote:
Originally Posted by mooninite
There is no 7950 GTX, however, there is a 7900 GTX. Are you sure what model of card you have?
Of course it exists. The PCI ID is 0x0297.
It's a mobile card, which is why probably why you couldn't find it.
http://us.download.nvidia.com/XFree8...ppendix-a.html
http://en.wikipedia.org/wiki/Compari...ce_Go_7_series

BTW - I've been running compiz all day and am still waiting for X to crash. It is flickering every now and then, ever since this showed up in the dmesg log:
[ 6542.199750] NVRM: Xid (0001:00): 6, PE0000 0200 07800000 0000f60c ff444444 00000001
[ 6542.216699] NVRM: Xid (0001:00): 36, L1 -> L0

This flickering always precedes X crashing - I can only assume they are related.
Attached is the bug report generated before the freeze, but after the flickering starts. I'll attach a second when X freezes.
Attached Files
File Type: log nvidia-bug-report.log (104.9 KB, 112 views)
pferraro is offline   Reply With Quote
Old 01-15-08, 10:11 PM   #22
pferraro
Registered User
 
Join Date: Aug 2007
Posts: 30
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

OK - so after several hours of annoying flickering, X finally freezes. As promised, here's the bug report taken by a remote ssh session.
The first Xid pair in dmesg marks the onset of the flickering:
[ 4860.462217] NVRM: Xid (0001:00): 6, PE0000 0200 07800000 0000ed0c 00000000 04b00000
[ 4860.479057] NVRM: Xid (0001:00): 36, L1 -> L0

The second Xid pair is the deadly one that freezes X:
[28745.291004] NVRM: Xid (0001:00): 6, PE0000 0200 07800000 0000ee64 ff444444 00000001
[28745.307536] NVRM: Xid (0001:00): 36, L0 -> L0
Attached Files
File Type: log nvidia-bug-report.log (87.0 KB, 97 views)
pferraro is offline   Reply With Quote
Old 01-22-08, 11:17 AM   #23
siarheig
Registered User
 
Join Date: Oct 2007
Posts: 1
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Did anyone try 169.09 to see if multi core bug is still there?
siarheig is offline   Reply With Quote
Old 01-22-08, 02:30 PM   #24
pferraro
Registered User
 
Join Date: Aug 2007
Posts: 30
Default Re: 169.07 BUG: The 7 series multi core bug still exsists.

Quote:
Originally Posted by siarheig
Did anyone try 169.09 to see if multi core bug is still there?
I see nothing in the change log that would indicate that this bug was fixed.
pferraro is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


Similar Threads
Thread Thread Starter Forum Replies Last Post
CPUMark99 - how do you compare fuelrod Benchmarking And Overclocking 66 07-19-11 08:32 AM

All times are GMT -5. The time now is 03:48 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.