Message boards :
Number crunching :
Heads up on NVIDIA driver 331.82
Message board moderation
Author | Message |
---|---|
kwalk Send message Joined: 2 Aug 99 Posts: 29 Credit: 16,535,464 RAC: 3 |
Installed 331.82 last night and this morning found all my CUDA tasks sitting going nowhere because of "missing GPUs". Backed off to 331.65 and all is fine again. Just a heads up to all you NVIDIA users. |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Were there any clues in the Event log you recall ? Had you shutdown Boinc or rebooted for example ? I ask because I'm bound to 331.76 dev driver for a couple of weeks, so can't really look at 331.82 yet. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Installed 331.82 last night and this morning found all my CUDA tasks sitting going nowhere because of "missing GPUs". Backed off to 331.65 and all is fine again. Just a heads up to all you NVIDIA users. You are the second occurance reported: No usable GPUs found after installing 7.2.28 I think the problem might be related to the newest nVidia driver released today. Because I noticed in the full log, that this log entry also appears once before I updated Boinc, and that's the only other thing changed. So it might be a problem with GeForce R331 Game Ready Driver 331.82 - WHQL Edit: Fixed link to post. Claggy |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14679 Credit: 200,643,578 RAC: 874 |
I've got the file downloaded, so I'll investigate when my current pair of GPUGrid tasks have finished (after lunch). |
kwalk Send message Joined: 2 Aug 99 Posts: 29 Credit: 16,535,464 RAC: 3 |
Were there any clues in the Event log you recall ? Had you shutdown Boinc or rebooted for example ?Like the novice that I am I failed to look at the logs. I shut down Boinc when ever I upgrade the drivers but all I looked at was that Boinc started back up after the upgrade. I did not look to see if it was "doing" anything until this morning. And of course I shut down Boinc and restarted it this morning when I rolled back the driver version. This was a pure and simple video driver upgrade then downgrade. No other changes were made, like no new version of Boinc etc. Sorry I could not be of more help. Edit: clarified up/down grade of video drivers only |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Were there any clues in the Event log you recall ? Had you shutdown Boinc or rebooted for example ?Like the novice that I am I failed to look at the logs. I shut down Boinc when ever I upgrade the drivers but all I looked at was that Boinc started back up after the upgrade. I did not look to see if it was "doing" anything until this morning. And of course I shut down Boinc and restarted it this morning when I rolled back the driver version. That's great, thanks :D. sounds like something to watch and I'm glad Claggy and Richard seem to be on the case here. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Don´t know if this could help. The missing GPU apears because the way the Geforce Experience install the new driver. Why, etc. make little sense to discuss. Good or Bad is the way NVidia choose to work. To fix the problem, follow the simple steps: - Before install the driver, totaly exit BOINC (not just the Boincmgr) - Install the driver normaly (don´t forget to choose clean instalation) - in some hosts you need to redo the instalation - the first one fails - nothing to be worry about, just redo the process, the second time it will work. - After the instalation ends, MANUALY restart the host so the windows could reload the drivers in the proper sequence, as needed to crunch. - Now all (games & crunching) will work fine. If you don´t do the Manual restart after the instalation, your driver will work for games for example but not for crunching, that´s the source of the Missing Gpu msg. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14679 Credit: 200,643,578 RAC: 874 |
Yes, definitely seems to be a problem - but the manual reboot is all that is needed to fix it. I had the 331.82-desktop-win8-win7-winvista-64bit-international-whql version downloaded manually, so bypassed the 'GeForce Experience' route. But I just did an 'express installation' over the top of the matching v331.65 - with BOINC closed down, of course. After restart, BOINC logged 20-Nov-2013 14:48:31 [---] [coproc] launching child process at D:\BOINC\\boinc.exe After system restart, BOINC started normally and detected the GPUs as normal. I'll bugrep to Rom - might be because we're now using NVCUDA.DLL 8.17.13.3182 NVIDIA CUDA 6.0.1 driver |
kwalk Send message Joined: 2 Aug 99 Posts: 29 Credit: 16,535,464 RAC: 3 |
We are on the right track. I did a clean install of 331.82 after shutting down Boinc. This time I rebooted the system before trying to start Boinc back up. Boinc is now happy as a clam with the 331.82 version of the drivers. So I guess the bottom line is stopping and starting Boinc is not enough. A system reboot is required. This time I remembered to grab the log file. I do not use the Gforce Experience to update the drivers if that is any help. Funny thing is I have never had to do this before so something has changed with the 331.82 version. 11/20/2013 10:03:39 AM | | cc_config.xml not found - using defaults 11/20/2013 10:03:39 AM | | Starting BOINC client version 7.2.28 for windows_x86_64 11/20/2013 10:03:39 AM | | log flags: file_xfer, sched_ops, task 11/20/2013 10:03:39 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 11/20/2013 10:03:39 AM | | Data directory: C:\ProgramData\BOINC 11/20/2013 10:03:39 AM | | Running under account kwalk 11/20/2013 10:03:39 AM | | CUDA: NVIDIA GPU 0: GeForce GTX 660 (driver version 331.82, CUDA version 6.0, compute capability 3.0, 2048MB, 1912MB available, 2132 GFLOPS peak) 11/20/2013 10:03:39 AM | | OpenCL: NVIDIA GPU 0: GeForce GTX 660 (driver version 331.82, device version OpenCL 1.1 CUDA, 2048MB, 1912MB available, 2132 GFLOPS peak) 11/20/2013 10:03:39 AM | | Host name: reaper 11/20/2013 10:03:39 AM | | Processor: 4 AuthenticAMD AMD Phenom(tm) II X4 940 Processor [Family 16 Model 4 Stepping 2] 11/20/2013 10:03:39 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni cx16 popcnt syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow 11/20/2013 10:03:39 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00) 11/20/2013 10:03:39 AM | | Memory: 4.00 GB physical, 8.00 GB virtual 11/20/2013 10:03:39 AM | | Disk: 931.51 GB total, 815.37 GB free 11/20/2013 10:03:39 AM | | Local time is UTC -5 hours 11/20/2013 10:03:39 AM | SETI@home | Found app_config.xml 11/20/2013 10:03:39 AM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 4257764; resource share 100 11/20/2013 10:03:39 AM | SETI@home | General prefs: from SETI@home (last modified 30-Jun-2013 22:04:18) 11/20/2013 10:03:39 AM | SETI@home | Computer location: home 11/20/2013 10:03:39 AM | SETI@home | General prefs: no separate prefs for home; using your defaults 11/20/2013 10:03:39 AM | | Reading preferences override file 11/20/2013 10:03:39 AM | | Preferences: 11/20/2013 10:03:39 AM | | max memory usage when active: 3071.53MB 11/20/2013 10:03:39 AM | | max memory usage when idle: 3685.83MB 11/20/2013 10:03:39 AM | | max disk usage: 100.00GB 11/20/2013 10:03:39 AM | | max CPUs used: 2 11/20/2013 10:03:39 AM | | (to change preferences, visit a project web site or select Preferences in the Manager) 11/20/2013 10:03:39 AM | | Not using a proxy 11/20/2013 10:03:41 AM | SETI@home | Restarting task 13se08ab.16132.16432.16.12.222_0 using setiathome_v7 version 700 in slot 0 11/20/2013 10:03:41 AM | SETI@home | Restarting task 18au09aa.23969.78586.9.12.127_1 using setiathome_v7 version 700 in slot 1 11/20/2013 10:03:41 AM | SETI@home | Restarting task 15no08ad.8637.481.14.12.171_0 using setiathome_v7 version 700 (cuda42) in slot 2 11/20/2013 10:03:41 AM | SETI@home | Restarting task 18au09aa.8636.67543.13.12.218_0 using setiathome_v7 version 700 (cuda42) in slot 3 11/20/2013 10:04:08 AM | SETI@home | Computation for task 15no08ad.8637.481.14.12.171_0 finished 11/20/2013 10:04:08 AM | SETI@home | Starting task 18au09aa.8636.67543.13.12.249_0 using setiathome_v7 version 700 (cuda42) in slot 2 |
Sp@ceNv@der Send message Joined: 10 Jul 05 Posts: 41 Credit: 117,366,167 RAC: 152 |
Had the same issue at home last night: but I noticed it shortly after installing the latest NVIDIA 331.82 drivers: normally I never upgrade these drivers with SETI/GPU-crunching active as it led to computation errors in the past. I also restart my machine completely after such an update, even it's not really needed anymore these days. Once all these things had been done, BOINC never had an issue with missing GPU(s) anymore. In my personal case, I've updated 3 crunchers with NVIDIA GPUs, without having to select "perform a clean installation" option: all went well on every machine, followed by a reboot. Kind regards To boldly crunch ... |
mr.mac52 Send message Joined: 18 Mar 03 Posts: 67 Credit: 245,882,461 RAC: 0 |
I had the same experience with 331.65 and a second clean install with manual reboot solved it for me. John |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Just a Tip: Always when install a new driver use the option: perform a clean installation That could save you a lot of headaches, you never know what the old driver leave behind. |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
This driver runs fine on my old Windows XP box but give me headache on my brand new notebook. About every 2-3 minutes i get driver restarts, downgrading to version 331.65 now... :/ Aloha, Uli |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
|
Paul D Harris Send message Joined: 1 Dec 99 Posts: 1122 Credit: 33,600,005 RAC: 0 |
shutdown boinc installed the new driver restart boinc no problems |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
This driver runs fine on my old Windows XP box but give me headache on my brand new notebook. About every 2-3 minutes i get driver restarts, downgrading to version 331.65 now... :/ Have to revise my own post, 331.82 now running perfect on Windows 8.1 64 bit. My problem were the settings in the "ap_cmdline_win_x86_SSE2_OpenCL_NV.txt" file, which were running smoothly on 331.65 but causing problems on 331.82. With an empty settings file the new driver runs fine now. There must have been changed a lot under the hood, more, the version number changing suggests. Aloha, Uli |
Wiggo Send message Joined: 24 Jan 00 Posts: 36828 Credit: 261,360,520 RAC: 489 |
I don't update video drivers unless I have a very good reason to do so, and it must be very good. Cheers. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
My problem were the settings in the "ap_cmdline_win_x86_SSE2_OpenCL_NV.txt" file, which were running smoothly on 331.65 but causing problems on 331.82. With an empty settings file the new driver runs fine now. Bad sign. Usually it means considerable performance degradation between driver versions. Did you check completion times for zero-blanked tasks? Did they change after upgrade? SETI apps news We're not gonna fight them. We're gonna transcend them. |
S@NL Etienne Dokkum Send message Joined: 11 Jun 99 Posts: 212 Credit: 43,822,095 RAC: 0 |
no problems on the Win7 box with a GTX660 in it. No noticable difference : running 2 AP's on the GPU - completion still in aprox. ~1.2 hours per WU (unless blanked of course) |
Tim Send message Joined: 19 May 99 Posts: 211 Credit: 278,575,259 RAC: 0 |
Install it to all my computers. None of them had problems. Not even the big one. Tim |
©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.