GTX 8800 and Cuda

Questions and Answers : GPU applications : GTX 8800 and Cuda
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · Next

AuthorMessage
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851526 - Posted: 10 Jan 2009, 2:12:21 UTC - in response to Message 851502.  

Ah - I was afraid of that - I can't register. They don't sent a validation email and I sent them an email to that effect but have yet to recieve a response.

BTW I really do appreciate all this help .... Thank you very much,
lou
ID: 851526 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851531 - Posted: 10 Jan 2009, 2:22:36 UTC - in response to Message 851526.  
Last modified: 10 Jan 2009, 2:23:17 UTC

This post has the updated exe CUDA app with the rest of the files needed in a zip file. I had the same problem with lunatics and had to go thru my spam filter to find their message to me.
ID: 851531 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851536 - Posted: 10 Jan 2009, 2:45:19 UTC - in response to Message 851531.  

Ok i went to geocites and got the file - I detached seti and then reattached halting all downloads. I stopped BOINC. I placed the downloaded files in the Seti directory and restarted BOINC and the files all ran for 28 sec CPU time and finished.

1/9/2009 6:40:32 PM||Starting BOINC client version 6.4.5 for windows_intelx86
1/9/2009 6:40:32 PM||log flags: task, file_xfer, sched_ops
1/9/2009 6:40:32 PM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
1/9/2009 6:40:32 PM||Data directory: C:\Program Files\BOINC\data
1/9/2009 6:40:32 PM||Running under account Louis
1/9/2009 6:40:32 PM|SETI@home|Found app_info.xml; using anonymous platform
1/9/2009 6:40:32 PM||Processor: 2 GenuineIntel Intel(R) Core(TM)2 CPU 6700 @ 2.66GHz [x86 Family 6 Model 15 Stepping 6]
1/9/2009 6:40:32 PM||Processor features: fpu tsc pae nx sse sse2 mmx
1/9/2009 6:40:32 PM||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
1/9/2009 6:40:32 PM||Memory: 2.00 GB physical, 5.84 GB virtual
1/9/2009 6:40:32 PM||Disk: 279.46 GB total, 147.37 GB free
1/9/2009 6:40:32 PM||Local time is UTC -8 hours
1/9/2009 6:40:32 PM||Not using a proxy
1/9/2009 6:40:32 PM||CUDA devices found
1/9/2009 6:40:32 PM||Coprocessor: GeForce 8800 GTX (1)
1/9/2009 6:40:32 PM|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 1783887; location: home; project prefs: default
1/9/2009 6:40:32 PM|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 3345687; location: (none); project prefs: default
1/9/2009 6:40:32 PM||General prefs: from SETI@home (last modified 08-Jan-2009 10:13:54)
1/9/2009 6:40:32 PM||Host location: none
1/9/2009 6:40:32 PM||General prefs: using your defaults
1/9/2009 6:40:32 PM||Preferences limit memory usage when active to 1023.23MB
1/9/2009 6:40:32 PM||Preferences limit memory usage when idle to 1841.82MB
1/9/2009 6:40:32 PM||Preferences limit disk usage to 2.00GB
1/9/2009 6:40:33 PM|Einstein@Home|Restarting task h1_0755.25_S5R4__297_S5R4a_2 using einstein_S5R4 version 610
1/9/2009 6:40:33 PM|SETI@home|Restarting task 06no08ab.7724.179226.4.8.22_2 using setiathome_enhanced version 605
1/9/2009 6:40:57 PM|SETI@home|Computation for task 06no08ab.7724.179226.4.8.22_2 finished
1/9/2009 6:40:57 PM|SETI@home|Starting 06no08aa.31393.16023.11.8.225_3
1/9/2009 6:40:57 PM|SETI@home|Starting task 06no08aa.31393.16023.11.8.225_3 using setiathome_enhanced version 605
1/9/2009 6:40:59 PM|SETI@home|Started upload of 06no08ab.7724.179226.4.8.22_2_0
1/9/2009 6:41:03 PM|SETI@home|Finished upload of 06no08ab.7724.179226.4.8.22_2_0
1/9/2009 6:41:35 PM|SETI@home|Computation for task 06no08aa.31393.16023.11.8.225_3 finished
1/9/2009 6:41:35 PM|SETI@home|Starting 01dc08ac.19772.12342.6.8.139_2
1/9/2009 6:41:35 PM|SETI@home|Starting task 01dc08ac.19772.12342.6.8.139_2 using setiathome_enhanced version 605
1/9/2009 6:41:37 PM|SETI@home|Started upload of 06no08aa.31393.16023.11.8.225_3_0
1/9/2009 6:41:39 PM|SETI@home|Finished upload of 06no08aa.31393.16023.11.8.225_3_0
1/9/2009 6:42:06 PM|SETI@home|Computation for task 01dc08ac.19772.12342.6.8.139_2 finished
1/9/2009 6:42:06 PM|SETI@home|Starting 01dc08ac.19772.12342.6.8.147_2
1/9/2009 6:42:06 PM|SETI@home|Starting task 01dc08ac.19772.12342.6.8.147_2 using setiathome_enhanced version 605
1/9/2009 6:42:08 PM|SETI@home|Started upload of 01dc08ac.19772.12342.6.8.139_2_0
1/9/2009 6:42:12 PM|SETI@home|Finished upload of 01dc08ac.19772.12342.6.8.139_2_0
1/9/2009 6:42:37 PM|SETI@home|Computation for task 01dc08ac.19772.12342.6.8.147_2 finished
1/9/2009 6:42:37 PM|SETI@home|Starting 05no08ab.8510.16432.10.8.223_2
1/9/2009 6:42:37 PM|SETI@home|Starting task 05no08ab.8510.16432.10.8.223_2 using setiathome_enhanced version 605
1/9/2009 6:42:39 PM|SETI@home|Started upload of 01dc08ac.19772.12342.6.8.147_2_0
1/9/2009 6:42:42 PM|SETI@home|Finished upload of 01dc08ac.19772.12342.6.8.147_2_0

The harder I works, the behinder I gets....
ID: 851536 · Report as offensive
Maik

Send message
Joined: 15 May 99
Posts: 163
Credit: 9,208,555
RAC: 0
Germany
Message 851540 - Posted: 10 Jan 2009, 3:02:18 UTC
Last modified: 10 Jan 2009, 3:04:24 UTC

just check your drivers.
go to device-manager ->gfx->properties->driver->driverdetails.
it should look like this:

from nvcpl.dll down to nvwssr.dll all should have version 6.14.11.8120
if not then drivers are not installed correctly. to do so, follow this instructions
ID: 851540 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851547 - Posted: 10 Jan 2009, 3:23:41 UTC - in response to Message 851540.  

All drivers are right on the money
ID: 851547 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851594 - Posted: 10 Jan 2009, 6:27:46 UTC - in response to Message 851547.  
Last modified: 10 Jan 2009, 6:38:05 UTC

All drivers are right on the money


Sorry, I couldn't find it if you already stated it, but what drivers are you using now and what have you tried?

There is available now for your card the new Beta 181.20 just recently released.

As well as the older Beta
180.48

And of course the standard 178.24
for your card.

I'd think the newer Beta drviers are the ones that might have the best results, but it seems to be differenet for each OS and pc.
You might have to try a few to find the one that works the best.
Looking at the results you're getting now, seems like a driver change is in order. I had the same problem using 180.48 but went to 180.60 on my card and made all the difference, so far haven't had a reason to go up to 181.20
ID: 851594 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851658 - Posted: 10 Jan 2009, 12:40:41 UTC - in response to Message 851594.  

Hi Byron,
I have tried 3 versions and am using the 181.20 right now. They all had the same effect. The device manager said they were properly installed.


The harder I works, the behinder I gets....
ID: 851658 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851671 - Posted: 10 Jan 2009, 13:40:19 UTC
Last modified: 10 Jan 2009, 14:19:35 UTC

Hi Byron,
I have tried 3 versions and am using the 181.20 right now. They all had the same effect. The device manager said they were properly installed.



I'm sure they were and are. The problem can be though not that the driver isn't installed correctly, but rather the CUDA app we have just doesn't like some of the drivers over others.

You have a common problem on WU true angle range is : 2.716657 and similar AR which probably are false positive -9 overflows.

You could use the method in this message to keep them off the server, and keep them from polluting the database with false results.

You can also use a batch file to spot Very Low Angle Range, or VLAR tasks (0.1 and lower anggle ranges ) that you can find examples of in this thread. Most likely VLAR tasks will be a problem to run and can also have side effects, like snow on the monitor and reboots, etc.

Also if you go back to the link you got the modified app file that I posted it's been updated to a newer version that may help with the overflows

You'll need to monitor your pc for a while till you're more aware what kind of angle ranges it can and can't run ( CUDA still a bit buggy even with the right drivers) once you determined you have a driver that is somewhat stable with the CUDA app, and take the precautions needed to avoid any problem ARs, it should be fairly smooth.

Out of the three drivers you've tried does this version seem the most stable to you?

One other thing, since we are now dealing with the modified CUDA app, it's been requested that we use the forum Modified SETI MB CUDA + opt AP package for full GPU utilization when giving help for that app so that the stock app can be discussed in the CUDA section.

Noticed one other thing on your card in the stderr. Not sure what it means but it seems to be different with most other people's cards I've looked at.
major = 1
minor = 0
seems like in the other cards I've looked at the minor=1 too. That may have a bearing on some of the problems you're having and you can raise that question in the other forum.
ID: 851671 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851679 - Posted: 10 Jan 2009, 14:19:11 UTC - in response to Message 851671.  

Thanks Byron but I am getting lost and confused. Is it possible for you to give me a step by step procedure for proper installation of BOINC?

This way we both know what is going on. I will start by detaching Seti. I will re-attach. All will be default.

Done and this is what I get - no cuda:

1/10/2009 6:17:53 AM||Starting BOINC client version 6.4.5 for windows_intelx86
1/10/2009 6:17:53 AM||log flags: task, file_xfer, sched_ops
1/10/2009 6:17:53 AM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
1/10/2009 6:17:53 AM||Data directory: C:\Program Files\BOINC\data
1/10/2009 6:17:53 AM||Running under account Louis
1/10/2009 6:17:53 AM||Processor: 2 GenuineIntel Intel(R) Core(TM)2 CPU 6700 @ 2.66GHz [x86 Family 6 Model 15 Stepping 6]
1/10/2009 6:17:53 AM||Processor features: fpu tsc pae nx sse sse2 mmx
1/10/2009 6:17:53 AM||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
1/10/2009 6:17:53 AM||Memory: 2.00 GB physical, 5.84 GB virtual
1/10/2009 6:17:53 AM||Disk: 279.46 GB total, 147.03 GB free
1/10/2009 6:17:53 AM||Local time is UTC -8 hours
1/10/2009 6:17:53 AM||Not using a proxy
1/10/2009 6:17:53 AM||CUDA devices found
1/10/2009 6:17:53 AM||Coprocessor: GeForce 8800 GTX (1)
1/10/2009 6:17:53 AM|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 1783887; location: home; project prefs: default
1/10/2009 6:17:53 AM|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 3345687; location: (none); project prefs: default
1/10/2009 6:17:53 AM||General prefs: from SETI@home (last modified 08-Jan-2009 10:13:54)
1/10/2009 6:17:53 AM||Host location: none
1/10/2009 6:17:53 AM||General prefs: using your defaults
1/10/2009 6:17:53 AM||Preferences limit memory usage when active to 1023.23MB
1/10/2009 6:17:53 AM||Preferences limit memory usage when idle to 1841.82MB
1/10/2009 6:17:53 AM||Preferences limit disk usage to 2.00GB
1/10/2009 6:17:54 AM|Einstein@Home|Restarting task h1_0755.25_S5R4__292_S5R4a_0 using einstein_S5R4 version 610
1/10/2009 6:17:54 AM|SETI@home|Restarting task 06no08ab.25418.180862.8.8.107_1 using setiathome_enhanced version 603


ID: 851679 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851686 - Posted: 10 Jan 2009, 14:29:22 UTC - in response to Message 851679.  
Last modified: 10 Jan 2009, 14:38:59 UTC

Thanks Byron but I am getting lost and confused. Is it possible for you to give me a step by step procedure for proper installation of BOINC?

This way we both know what is going on. I will start by detaching Seti. I will re-attach. All will be default.

Done and this is what I get - no cuda:

1/10/2009 6:17:53 AM||Starting BOINC client version 6.4.5 for windows_intelx86
1/10/2009 6:17:53 AM||log flags: task, file_xfer, sched_ops
1/10/2009 6:17:53 AM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
1/10/2009 6:17:53 AM||Data directory: C:\Program Files\BOINC\data
1/10/2009 6:17:53 AM||Running under account Louis
1/10/2009 6:17:53 AM||Processor: 2 GenuineIntel Intel(R) Core(TM)2 CPU 6700 @ 2.66GHz [x86 Family 6 Model 15 Stepping 6]
1/10/2009 6:17:53 AM||Processor features: fpu tsc pae nx sse sse2 mmx
1/10/2009 6:17:53 AM||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
1/10/2009 6:17:53 AM||Memory: 2.00 GB physical, 5.84 GB virtual
1/10/2009 6:17:53 AM||Disk: 279.46 GB total, 147.03 GB free
1/10/2009 6:17:53 AM||Local time is UTC -8 hours
1/10/2009 6:17:53 AM||Not using a proxy
1/10/2009 6:17:53 AM||CUDA devices found
1/10/2009 6:17:53 AM||Coprocessor: GeForce 8800 GTX (1)

1/10/2009 6:17:53 AM|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 1783887; location: home; project prefs: default
1/10/2009 6:17:53 AM|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 3345687; location: (none); project prefs: default
1/10/2009 6:17:53 AM||General prefs: from SETI@home (last modified 08-Jan-2009 10:13:54)
1/10/2009 6:17:53 AM||Host location: none
1/10/2009 6:17:53 AM||General prefs: using your defaults
1/10/2009 6:17:53 AM||Preferences limit memory usage when active to 1023.23MB
1/10/2009 6:17:53 AM||Preferences limit memory usage when idle to 1841.82MB
1/10/2009 6:17:53 AM||Preferences limit disk usage to 2.00GB
1/10/2009 6:17:54 AM|Einstein@Home|Restarting task h1_0755.25_S5R4__292_S5R4a_0 using einstein_S5R4 version 610
1/10/2009 6:17:54 AM|SETI@home|Restarting task 06no08ab.25418.180862.8.8.107_1 using setiathome_enhanced version 603



Why do you want to reinstall Boinc? The line in red shows you have CUDA.

Just go back to the opti app and download the mod updated CUDA app and give it another try or continue to use the stock. In my opinion the mod app is easier on your system.

If you need we can go over the steps in my last message and there are also others at the forum link I gave that may be able to explain it better than me. I'm proabably not the best communicator here. It is important to realize though that CUDA still is a Beta app, even the stock, so even under the best circumstances it still has problems and is difficult to run.
ID: 851686 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851692 - Posted: 10 Jan 2009, 14:44:27 UTC - in response to Message 851686.  

I did not reinstall BOINC - I just detached from seti and reattached.
Getting back to basics - I have downloaded the file opti.zip. I plan to:

seti is not taking any new tasks so I will abort the three seti tasks no in line on my computer and update seti to clear them.

Then I will stop BOINC, extract all files from opti.zip and place them in the /projects/seti... folder - replacing the existing files.

then restart BOINC. Is this correct?

Another thing, I have a 280 card on another machine running Vista. It is fine. I have my Seti preferences set to no AP's and download a full days work. The vista machine heeds these. So this machine should do so also. But it doesn't. It will download 3 WU's only when using stock app and usually 4 with the optimized package. Question is should I place the opti files in the vista machine too?
ID: 851692 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851695 - Posted: 10 Jan 2009, 14:54:34 UTC - in response to Message 851692.  
Last modified: 10 Jan 2009, 14:59:04 UTC

seti is not taking any new tasks so I will abort the three seti tasks no in line on my computer and update seti to clear them.
If you still have Seti tasks waiting to run you don't need to abort them. You can just install the app and process them. The Duration Correction Factor was reset when you detached and will take a while to increase, but will in time. You won't run out of work, but it may wait till you've completed all the tasks you have before assigning new work for a while.

Then I will stop BOINC, extract all files from opti.zip and place them in the /projects/seti... folder - replacing the existing files.
then restart BOINC. Is this correct?

Yes

Another thing, I have a 280 card on another machine running Vista. It is fine.

The 8800 series has been know to have problems with Seti Cuda and compute campability. some have 1.1 others only 1.0. This could be part of the problem but the mod app has been known to help with this.

Question is should I place the opti files in the vista machine too?

It's more optimized than the stock app so it should be somewhat faster for that card too. It should also make it easeir for it to do the tasks too since that's what the code in it is designed to do. It's the same basic code as the stock with some tweaking.
ID: 851695 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851696 - Posted: 10 Jan 2009, 15:04:43 UTC - in response to Message 851695.  

Done and it added cuda the the application name.

It downloaded 2 more tasks and immediately finished the one it was working on and started the second. The 2nd is running now but....as usual CPU time goes to 27 sec and stops increasing. The time for completion for this task is not increasing. It has done this in the past and I halted the task after 5 hours or running.

Damn

Going to put the app on my Vista machine...
ID: 851696 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851699 - Posted: 10 Jan 2009, 15:11:48 UTC - in response to Message 851696.  

Ok, So you're getting an overflow in angle range 0.378811. That seems to be a problem for the driver you have for the card. 0.4xxxx also seems to be a no go either as well as 2.7 when it comes to overflows.

From what you remember from the past drivers you used, are you seeing less errors and problems with your computer (like snow on the screen, reboots, freezing, etc ) with these drivers than with the others. Or has the main problem always been just the overflows?
ID: 851699 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851701 - Posted: 10 Jan 2009, 15:20:53 UTC - in response to Message 851699.  

It has always been the same... they either complete in a few minutes asking for .01 to .08 credits or run forever. The driver did not seem to make any difference. Should I have downloaded the v.20 cuda dribers from nvidia?
ID: 851701 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851706 - Posted: 10 Jan 2009, 15:33:02 UTC - in response to Message 851701.  
Last modified: 10 Jan 2009, 15:54:02 UTC

It has always been the same... they either complete in a few minutes asking for .01 to .08 credits or run forever. The driver did not seem to make any difference. Should I have downloaded the v.20 cuda dribers from nvidia?

To be honest with you, you need to download any drivers from 178.24 to 181.xx and try them all. And then try them again. In most cases there's only one that works the best. There will still be overflows in AR 2.7 and some others, but down to like one or two every 10 successfull ones or so.

You should reboot the pc once you have one of the 27 sec tasks (overflows), that has been known to clear it up at least for a while. Of course if you keep rebooting and it just keeps happening, then it's time to try another driver if there are any left.

And most likely the VLAR tasks will either compute error, or stall and have to be aborted, or could even freeze or BSOD your pc (that's why we abort them) so you'll want to run the batch file I mentioned earlier to keep those out.

Edit: I checked again on the nvidia site and the 181.20 are the drivers they recommend as the current CUDA driver for your 8800GTX

I also just got some feed back on what I mentioned I had seen in the stderr of the tasks you were doing and it looks like you do have the 1.0 compute capability with that card. Seti works on 1.1 so there's a chance the card just isn't going to work with the app in it's current state.

Question, what does

major = 1
minor = 0

or

major = 1
minor = 1

in the stderr stand for?



Think revision version of CUDA (so called computational ability or smth like this - check correct spelling on nVidia site).
In general 1.0 - 8800 cards. Lack of some very useful abilities. All with 1.1 and up have mostly stuff that needed fore real work.
ID: 851706 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851726 - Posted: 10 Jan 2009, 16:08:39 UTC - in response to Message 851671.  

Damn - what am I to do?

I just installed 178_20 drivers - seti downloaded many wu's and completed 3 in 28 seconds - I then suspended seti but I sure could use some simple instructions. It is getting back into the confused state.

I have no idea what the stdrr stands for. It has been a standard installation with the exception of the opti package
ID: 851726 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 851730 - Posted: 10 Jan 2009, 16:12:34 UTC


Compute capability can be checkedd with a utility CUDA-Z.

I read somewhere that a compute capability of 1.1 is required to use Cuda on Boinc/Seti MultiBeam.
Boinc....Boinc....Boinc....Boinc....
ID: 851730 · Report as offensive
Profile Byron S Goodgame
Volunteer tester
Avatar

Send message
Joined: 16 Jan 06
Posts: 1145
Credit: 3,936,993
RAC: 0
United States
Message 851743 - Posted: 10 Jan 2009, 16:30:22 UTC - in response to Message 851726.  
Last modified: 10 Jan 2009, 16:33:10 UTC

Damn - what am I to do?

Pretty much what you're doing. Trying different drivers to see if one is more consistant than the other, and hopefully one that can run the tasks without resulting in just overflows

I just installed 178_20 drivers - seti downloaded many wu's and completed 3 in 28 seconds - I then suspended seti but I sure could use some simple instructions. It is getting back into the confused state.

It's really just a matter of wether you can find a driver that will run at this point. There have been others I've seen here with 1.0 that have gotten it to work, though I can't say they didn't end up having the same problem in the end.

You will want to keep an eye on the different angle ranges that you will see in the stderr file, so you can see if there are any angle ranges that you are able to complete. It's not uncommon for there to be angle ranges that will create overflows and errors. It's only a problem when you can't do any AR's.

I have no idea what the stdrr stands for. It has been a standard installation with the exception of the opti package

I'm sorry, this is the stderr file. Each task has one and you can access them through the Tasks for host page by clicking on the Task ID. Before the task is completed there isn't much info, but after it's completed and back on the server you can access the task details from it.
ID: 851743 · Report as offensive
Profile The Mad Banger
Volunteer tester
Avatar

Send message
Joined: 6 May 07
Posts: 26
Credit: 577,038
RAC: 0
United States
Message 851746 - Posted: 10 Jan 2009, 16:32:49 UTC - in response to Message 851730.  

cuda-z says my compute capability is 1.0 - does that mean I have to stop using cuda for seti?
ID: 851746 · Report as offensive
Previous · 1 · 2 · 3 · Next

Questions and Answers : GPU applications : GTX 8800 and Cuda


 
©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.