Page 1 of 1
GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Sep 15, 2012 11:37 am
by warp-9.9
Can anyone else confirm this? System is running Win7 64-bit. System ran fine on the old 301.24 and 301.42 drivers. System runs fine when doing other things now, but consistently hangs within 15 seconds of starting Folding@Home, and BSOD 5s afterwards.
It's either my recent GeForce 306.23 driver update for my GTX480 boards, or a recent BIOS update for my ASUS Sabertooth 990FX (R1.0), BIOS 1304, or a BIOS setting (but I think everything is the same).
These are the only two pieces of software that have changed.
Update: If I hurry and pause the GPU slot after start, then the system processes the SMP slot just fine and seems to remain stable.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Sep 15, 2012 11:51 am
by bollix47
FWIW
I just downloaded 306.23 for my GTX 480 running on Win 7 64-bit yesterday and the client has completed a WU with no problems noticed so far but it's a bit early.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Sep 15, 2012 12:44 pm
by heikosch
No problems with GeForce 306.23 and GTX680, GTX570 and GTX460.
Heiko
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Thu Dec 13, 2012 2:13 pm
by Radiohead
Oddly enough I have 2 460GTXs and each one of them BSOD with the 306 drivers after clean install if I open a web browser (either IE or Moz) when the client is running. Windows error reporting shows same result and shows same error in event viewer. I'm getting frustrated because the only time these two cards won't crap out is when I'm running an old 2 series driver. Wondering if it has something to do with the HD audio driver most of these cards require for HDMI?
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Thu Dec 13, 2012 3:18 pm
by P5-133XL
If you have a driver that doesn't cause you problems -- Just use it. The basic theory is: If it isn't broke, then don't fix it for all you can do is harm.
Something is going on if you are getting blue-screens that needs to be addressed. That isn't normal for any driver and folding should not produce them if there are not other issues (like the video card is over heating or is OC'ed).
The changes in recent drivers for the most part are modifications for Keplers. Your GTX 460's are too old to benefit from the vast majority of changes in modern drivers. I like 260.99 (for my GTX 460's) mainly because lag seems to be pretty much absent as compared to more recent drivers. I don't notice any benefits going newer and I have tried. That being said, I'm not getting blue-screens with newer drivers either.
My guess is that you are having issues with browsers (while folding) because most browsers now use their GPU to render videos, flash, and other components of web pages. Try disabling those features in the browser you use. Personally, I use Chrome and in the shortcut I appended --disable-gpu and that works very well.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sun Dec 16, 2012 6:20 pm
by Jimboc
P5-133XL wrote:If you have a driver that doesn't cause you problems -- Just use it. The basic theory is: If it isn't broke, then don't fix it for all you can do is harm.
Something is going on if you are getting blue-screens that needs to be addressed. That isn't normal for any driver and folding should not produce them if there are not other issues (like the video card is over heating or is OC'ed).
My guess is that you are having issues with browsers (while folding) because most browsers now use their GPU to render videos, flash, and other components of web pages. Try disabling those features in the browser you use. Personally, I use Chrome and in the shortcut I appended --disable-gpu and that works very well.
Hi Radiohead,
Sorry for not posting sooner but I have been using the 306.23 Nvidia driver since its release in September 2012 with my 2x EVGA Geforce GTX 580 SC and I have not experienced a single issue. I am using my GTX 580s for folding on a very regular basis.
I am still using the 306.23 driver since the newer 310 drivers are causing too many issues with recent games (according to the Nvidia forums). I have not installed the newer driver yet but will do so when the 310 driver has improved in this area.
I agree with P5-133XL that the BSODs are being caused by something else that you will need to determine what that is. His suggestion of disabling GPU acceleration for your internet browser is a good start to troubleshooting this issue. Please note that I don’t have the bundled HD Audio driver installed on my PC since I don’t use the HD Audio capability.
I hope this helps. Thank you.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Wed Dec 19, 2012 5:11 pm
by rhavern
I used 266 for a very long time until beta testing a new core_16 required 270 or newer to work. I then used 285 for quite a while until an MS update snuck in 306.97. As it hasn't caused any problems and all my GPU folders are dedicated, I haven't bothered to change back to 285. This is across G92 and Fermi GPUs. YMMV
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Jan 12, 2013 8:23 pm
by warp-9.9
ASUS Sabertooth mainboard BIOS has had several updates over the past few months which state "Increase CPU stability" as one of the changes. I speculate that they may have made a change that caused problems, and tried to fix it. That would explain why the F@H on the CPU (AMD FX-8150) was failing, and along with driver interaction, may have accounted for the BSODs in my case.
I ran the interim BIOSes from 1304 to the current 1604, along with the video drivers up to the current 310.90, and the BSODs decreased and have now totally stopped. The occurrence of failed Work Units (on CPU only) decreased slightly, but I do still get failed work units (on CPU only) sometimes several in one day, sometimes I go 2-3 days without a fail. It is far to frequent to be useful or reliable, and mostly is just wasting CPU cycles. The BIOS upgrades helped some other areas at the expense of breaking CPU functionality for over 6 months now since that initial broken BIOS, and updates are fewer and farther between, indicating the end of life cycle draws near, and there's little hope of further updates from ASUS.
On the other hand, the 310.90 drivers have been running very stable and smooth the past few days, and actually allow me to manipulate windows without much lag (tolerable and functional Windowing). I can also get only mild to moderate stuttering during video playback or video games even in windowed modes. Finally, the GPUs seem to be running almost 10C cooler. I am wondering if they tweaked the GPU driver to handle interrupts better, and throttle back the usage under "full load" conditions, to leave a little overhead for multiple apps that may come along asking to use the GPU.
Never hurts to upgrade or downgrade, as it can always be reverted. Just a matter of finding which combination of BIOSes, firmwares and drivers achieve what you want: disk performance, CPU stability, and both for GPU.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Jan 12, 2013 8:58 pm
by art_l_j_PlanetAMD64
If you are getting so many failed CPU/SMP WUs, it indicates that there a problem with your computer. It may be an overly-aggressive CPU and/or memory overclock (if you are overclocking), or some component (like memory or the PSU) that is marginal, or even a bad contact in a cable somewhere, but there should not be so many failed CPU/SMP WUs. My 5 computers here have completed hundreds of SMP WUs without experiencing a failed WU.
It would be a good idea to run some health-check programs, like memtest86, and stress-test programs like Prime95, until you find out the cause of the instability in your computer. It should be an extremely rare occurrence to have a failure of a CPU/SMP WU.
Re: GeForce 306.23 may be causing BSOD on GTX480
Posted: Sat Jan 12, 2013 9:55 pm
by JimF
Radiohead wrote:Oddly enough I have 2 460GTXs and each one of them BSOD with the 306 drivers after clean install if I open a web browser (either IE or Moz) when the client is running.
Some versions of Firefox have caused BSODs for me. I have solved the problem by setting the memory timings in the motherboard back to their default value (e.g., 5-5-5-15 for this P45 board), even though the DDR2 memory modules are rated for 4-4-4-12. But other versions of Firefox are more stable, so YMMV.