video driver crash while crunching Seti

Message boards : Number crunching : video driver crash while crunching Seti
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1314630 - Posted: 13 Dec 2012, 13:01:57 UTC
Last modified: 13 Dec 2012, 13:24:46 UTC

I'm using x41g under Windows 7 Pro x64 and driver 306.97 WHQL, Boinc 6.10.58 x64 and Boinctasks 1.44 x64, I've also noticed this behavior under x41u and under x41zb, of course with x41zb I've encountered a BSOD which was video driver related, I've recently, cleaned the PC and the aircooled GTX590 of dust, the PC now has 8 Koolance 108cfm fans, 2 of these fans are for the cpu only as their in push/pull for the Corsair H70 Radiator(the psu is a Silverstone ST1500 1500w power supply and the 590 is getting 50A straight from the psu and as much as it can use from the pcie slot), temps are usually around 68C, except when the heater is on, then it's about 73C, while I was at it I added an EVGA power booster x1 pcie card to deliver clean 12v power to the pcie slots just like the Rampage motherboards do. the driver crash really just effects the fan, it drops from 100% to somewhere around 39-49%, except for that BSOD that is, but that BSOD may have been cause of dust which has since been cleaned out. I just went back to x41zb since I've seen crashes under all 3 versions and I doubt any of them are the cause, but then I don't have a spare EVGA GTX590 to test whether it's the video card that I have now or the GTX590's cards in particular.

1877 SETI@home 12/13/2012 3:10:22 AM Starting task 24dc09ab.16688.8815.140733193388040.10.69_0 using setiathome_enhanced version 610
1878 SETI@home 12/13/2012 3:10:22 AM Starting 24dc09ab.29515.5543.140733193388041.10.48_1
1879 SETI@home 12/13/2012 3:10:22 AM Starting task 24dc09ab.29515.5543.140733193388041.10.48_1 using setiathome_enhanced version 610
1880 SETI@home 12/13/2012 3:10:22 AM Starting 24dc09ab.29515.5543.140733193388041.10.16_0
1881 SETI@home 12/13/2012 3:10:22 AM Starting task 24dc09ab.29515.5543.140733193388041.10.16_0 using setiathome_enhanced version 610
1882 SETI@home 12/13/2012 3:10:22 AM Starting 24dc09ab.29515.5543.140733193388041.10.10_1
1883 SETI@home 12/13/2012 3:10:22 AM Starting task 24dc09ab.29515.5543.140733193388041.10.10_1 using setiathome_enhanced version 610
1884 SETI@home 12/13/2012 3:12:42 AM Computation for task 24dc09ab.16688.8815.140733193388040.10.69_0 finished
1885 SETI@home 12/13/2012 3:12:42 AM Computation for task 24dc09ab.29515.5543.140733193388041.10.48_1 finished
1886 SETI@home 12/13/2012 3:12:42 AM Computation for task 24dc09ab.29515.5543.140733193388041.10.10_1 finished
1887 SETI@home 12/13/2012 3:12:43 AM Restarting task 24ap10aa.15933.13977.140733193388045.10.21_1 using setiathome_enhanced version 610
1888 SETI@home 12/13/2012 3:12:43 AM Restarting task 23fe11aa.27679.5388.140733193388047.10.113_1 using setiathome_enhanced version 610
1889 SETI@home 12/13/2012 3:12:43 AM Restarting task 23fe11aa.27679.6206.140733193388047.10.198_0 using setiathome_enhanced version 610
1890 SETI@home 12/13/2012 3:12:43 AM Restarting task 23fe11aa.27679.6206.140733193388047.10.115_0 using setiathome_enhanced version 610

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1314630 · Report as offensive
Profile Cliff Harding
Volunteer tester
Avatar

Send message
Joined: 18 Aug 99
Posts: 1432
Credit: 110,967,840
RAC: 67
United States
Message 1314667 - Posted: 13 Dec 2012, 15:38:31 UTC

I have noticed that if you have Precision X running minimized, it will control the fan speed in a more effecient manner. Noticed this after installing my GTX660SC. It seems that the fan speed was not being properly regulated and the card began to overheat and went BSOD. This happened immediately after installation and doing a clean install of 306.97.

Also under the FAN tab in Profiles area, make sure you check FORCE FAN SPEED UPDATE EACH PERIOD. Currently using Precision X 3.0.04, which is controlling 1 x GTX660SC & 1x GTX460SE on the i7/950. Fan speed has been on a average of 65c for the 660SC and 53-55 for the 460SE. This has been so effective that it is also controlling 1 x GTX275/GTS250 & 1 X GTS250 on the i7/930.

Currently running on the i7/950 (Win7 x64, Nvidia 310.70(beta), BOINC 7.0.42, Lunatics _x41g).


I don't buy computers, I build them!!
ID: 1314667 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1314674 - Posted: 13 Dec 2012, 16:14:12 UTC - in response to Message 1314667.  
Last modified: 13 Dec 2012, 16:19:27 UTC

I'm using the MSI Afterburner program, v2.2.4 and the fan is at 100%.

I'll try 3.0.4.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1314674 · Report as offensive
Profile Cliff Harding
Volunteer tester
Avatar

Send message
Joined: 18 Aug 99
Posts: 1432
Credit: 110,967,840
RAC: 67
United States
Message 1314675 - Posted: 13 Dec 2012, 16:25:12 UTC - in response to Message 1314674.  

I'm using the MSI Afterburner program, v2.2.4 and the fan is at 100%.

I'll try 3.0.4.


Are you using Afterburner to control the fan is a manual or automatic mode?


I don't buy computers, I build them!!
ID: 1314675 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1314680 - Posted: 13 Dec 2012, 16:44:33 UTC - in response to Message 1314675.  

I'm using the MSI Afterburner program, v2.2.4 and the fan is at 100%.

I'll try 3.0.4.


Are you using Afterburner to control the fan is a manual or automatic mode?

Automatic, force won't appear without it, fan set to 100%.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1314680 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315024 - Posted: 14 Dec 2012, 6:58:03 UTC
Last modified: 14 Dec 2012, 7:00:18 UTC

Alright I just had a video driver crash at 10:51pm, I'd been crunching with x41zb since 7:00pm and I'm using EVGA Precision X 3.0.4 and I have force and automatic on and the fan speed went from 100% to 49% and it's now at 68% and the fan has not gone back to 100%, the fan is now at 67%, 66%, 64%, 66%, 65%, 64%, 66%...


Fan speed was not on auto, so I manually put it there and checked the little box for auto.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315024 · Report as offensive
Horacio

Send message
Joined: 14 Jan 00
Posts: 536
Credit: 75,967,266
RAC: 0
Argentina
Message 1315110 - Posted: 14 Dec 2012, 14:59:10 UTC

Ive seen that after Windows have recovered from a crashed driver, neither Precision, neither Afterburner are able to handle the GPU anymore without a reboot... restarting just the OC utilities didnt helped neither... (Im not using the very last version of this apps, but until now, all of them have had this behaviour)

About the crashes, there is some insight in the readme of the x41zb saying that certain values for the optional pulsefindings parameters in the mbcuda.cfg could lead to drive errors... may be you need to reduce those values a bit?
ID: 1315110 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315124 - Posted: 14 Dec 2012, 15:36:57 UTC - in response to Message 1315110.  

Ive seen that after Windows have recovered from a crashed driver, neither Precision, neither Afterburner are able to handle the GPU anymore without a reboot... restarting just the OC utilities didn't helped neither... (I'm not using the very last version of this apps, but until now, all of them have had this behaviour)

About the crashes, there is some insight in the readme of the x41zb saying that certain values for the optional pulsefindings parameters in the mbcuda.cfg could lead to drive errors... may be you need to reduce those values a bit?

After I'd set auto on the checkbox for Precision X the fan speed stayed at 100% for the rest of the night, this software is complicated or at least that is the way that this software appears to be to Me. I'll see tonight if that final adjustment does any good or not... If not I'd wish they'd fix that as that seems like a bug to Me.

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315124 · Report as offensive
Horacio

Send message
Joined: 14 Jan 00
Posts: 536
Credit: 75,967,266
RAC: 0
Argentina
Message 1315162 - Posted: 14 Dec 2012, 17:45:08 UTC - in response to Message 1315124.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an aditional config file (mbcuda.cfg) to set certain advanced options...
ID: 1315162 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315170 - Posted: 14 Dec 2012, 18:06:27 UTC - in response to Message 1315162.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an additional config file (mbcuda.cfg) to set certain advanced options...

I'm guessing that the cfg file after being edited goes in the same folder as x41zb? That's where I put My copy.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315170 · Report as offensive
Horacio

Send message
Joined: 14 Jan 00
Posts: 536
Credit: 75,967,266
RAC: 0
Argentina
Message 1315292 - Posted: 14 Dec 2012, 22:51:50 UTC - in response to Message 1315170.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an additional config file (mbcuda.cfg) to set certain advanced options...

I'm guessing that the cfg file after being edited goes in the same folder as x41zb? That's where I put My copy.

Yes, in the project directory, and it should be referenced in the app_info.xml so BOINC can keep it there and/or copy it to the slots as needed...
(the right way to add the references in the app_info can be seen in the .aistub file that comes in packs)
ID: 1315292 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315296 - Posted: 14 Dec 2012, 23:00:43 UTC - in response to Message 1315292.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an additional config file (mbcuda.cfg) to set certain advanced options...

I'm guessing that the cfg file after being edited goes in the same folder as x41zb? That's where I put My copy.

Yes, in the project directory, and it should be referenced in the app_info.xml so BOINC can keep it there and/or copy it to the slots as needed...
(the right way to add the references in the app_info can be seen in the .aistub file that comes in packs)

I figured that out just as the forum took a nap... :o
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315296 · Report as offensive
Profile Cliff Harding
Volunteer tester
Avatar

Send message
Joined: 18 Aug 99
Posts: 1432
Credit: 110,967,840
RAC: 67
United States
Message 1315382 - Posted: 15 Dec 2012, 1:56:07 UTC - in response to Message 1315124.  
Last modified: 15 Dec 2012, 1:57:39 UTC

After I'd set auto on the checkbox for Precision X the fan speed stayed at 100% for the rest of the night, this software is complicated or at least that is the way that this software appears to be to Me. I'll see tonight if that final adjustment does any good or not... If not I'd wish they'd fix that as that seems like a bug to Me.


Click on the two gears (Properties) in the upper right corner of Precision X and click on the fan tab. Make sure that the boxes for ENABLE SOFTWARE AUTOMATIC FAN CONTROL and FORCE FAN SPEED UPDATE ON EACH PERIOD are both checked. Make sure that WINDOWS STARTUP is check on the main panel. Minimize the application to the notification area. Never had a problem even after a video driver crash.


I don't buy computers, I build them!!
ID: 1315382 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315384 - Posted: 15 Dec 2012, 2:01:30 UTC - in response to Message 1315382.  

After I'd set auto on the checkbox for Precision X the fan speed stayed at 100% for the rest of the night, this software is complicated or at least that is the way that this software appears to be to Me. I'll see tonight if that final adjustment does any good or not... If not I'd wish they'd fix that as that seems like a bug to Me.


Click on the two gears (Properties) in the upper right corner of Precision X and click on the fan tab. Make sure that the boxes for ENABLE SOFTWARE AUTOMATIC FAN CONTROL and FORCE FAN SPEED UPDATE ON EACH PERIOD are both checked. Make sure that WINDOWS STARTUP is check on the main panel. Minimize the application to the notification area. Never had a problem even after a video driver crash.

Their all checked and yeah I double checked the boxes, their smashed.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315384 · Report as offensive
Profile arkayn
Volunteer tester
Avatar

Send message
Joined: 14 May 99
Posts: 4438
Credit: 55,006,323
RAC: 0
United States
Message 1315420 - Posted: 15 Dec 2012, 3:49:17 UTC - in response to Message 1315292.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an additional config file (mbcuda.cfg) to set certain advanced options...

I'm guessing that the cfg file after being edited goes in the same folder as x41zb? That's where I put My copy.

Yes, in the project directory, and it should be referenced in the app_info.xml so BOINC can keep it there and/or copy it to the slots as needed...
(the right way to add the references in the app_info can be seen in the .aistub file that comes in packs)


I don't have it referenced in my app_info, so it is not absolutely necessary.

ID: 1315420 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315427 - Posted: 15 Dec 2012, 4:01:03 UTC - in response to Message 1315420.  

That could be, but I don't see how, as this is the Windows version, their just a precompiled exe file and 2 dll files, nothing to configure here.

Ive got the apps, from this link posted by Jason here...
Those are all Windows versions and they use an additional config file (mbcuda.cfg) to set certain advanced options...

I'm guessing that the cfg file after being edited goes in the same folder as x41zb? That's where I put My copy.

Yes, in the project directory, and it should be referenced in the app_info.xml so BOINC can keep it there and/or copy it to the slots as needed...
(the right way to add the references in the app_info can be seen in the .aistub file that comes in packs)


I don't have it referenced in my app_info, so it is not absolutely necessary.

I put it in anyway, as it can't hurt.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315427 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1315465 - Posted: 15 Dec 2012, 7:18:19 UTC

I had a crash again, but I'm making progress, the fan stayed at 100%, now I think I set the card with mbcuda.cfg to where it won't crash again, I hope. The following is what I set it at:

processpriority = belownormal
pfblockspersm = 7
pfperiodsperlaunch = 100


I'd set x41zb like this just before the crash:

processpriority = belownormal
pfblockspersm = 9
pfperiodsperlaunch = 150


Needless to say I put settings back to the first example.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1315465 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65689
Credit: 55,293,173
RAC: 49
United States
Message 1318673 - Posted: 22 Dec 2012, 6:41:38 UTC
Last modified: 22 Dec 2012, 6:42:28 UTC

Ok I'm not sure why, but on the EVGA GTX590 Classified(a 1598) gpu2 is at 0.938v and gpu1 is at 0.963v, the fan is on gpu2's panel which is strange, If I put both gpus at 0.938v the driver crashes while doing Seti, yet if the card is left as it is now, no crashes, the card works at least. So as long as I keep the filters clean and such I should have no problems, I'm using Precision X 3.0.4, BoincTasks 1.44 x64, Boinc 6.10.58 x64 under Windows 7 Pro x64 and of course x41zc with the following set in the mbcuda.cfg file. Currently the temperature for both gpus is at 69-70C and the fan is at 100%.

processpriority = belownormal
pfblockspersm = 14
pfperiodsperlaunch = 100

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 1318673 · Report as offensive

Message boards : Number crunching : video driver crash while crunching Seti


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.