Message boards :
Number crunching :
Can't update or reset project. TIMEOUT
Message board moderation
Author | Message |
---|---|
woodyrox Send message Joined: 7 Apr 01 Posts: 34 Credit: 16,069,169 RAC: 0 |
I mostly run SETI when temperatures get cold and I can use the waste heat to help heat the house. So I'm starting SETI tasks on my computers and am having problems with one machine. Last spring, it was running Lunatics, but when I tried to get tasks, the system said I didn't have a valid SETI app. So I uninstalled Lunatics and installed the original SETI software. I've tried several versions, and they all behave the same. Scheduler requests timeout. Resetting has no apparent effect. I was able to report completed tasks once, but that's it. I've tried detaching and reattaching. No significant improvement. I'm able to get CUDA-FERMI tasks for the GPU, completed results upload, but I can't report them. I detached, lost a bunch of completed work units re-attached, got some new work units, but again can't report them. All my other computers work fine. In fact, they are real snappy when updating. They are still running the old Lunatics version. One more thing, in the process of bringing this machine on task, my NVIDIA GT240 graphic card fan failed, so I replaced it with a GT430. The drivers recognized the card & GPU, SETI loaded the appropriate CUDA-Fermi app, and the GPU is crunching tasks and completing without errors. I have yet to get a CPU task, and the CPU hasn't changed. Here's what is logged: Fri 09 Nov 2012 11:33:00 AM EST Project communication failed: attempting access to reference site Fri 09 Nov 2012 11:33:00 AM EST SETI@home Scheduler request failed: Timeout was reached Fri 09 Nov 2012 11:33:01 AM EST Internet access OK - project servers may be temporarily down. The problem computer is here: http://setiathome.berkeley.edu/results.php?hostid=6109088 If anyone has any ideas where I should look for problems, help would be greatly appreciated. Thanks |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Most likely not a problem on your end. There have been server availability problems for the last week or so. Peruse the Panic Mode thread for more information. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Fred J. Verster Send message Joined: 21 Apr 04 Posts: 3252 Credit: 31,903,643 RAC: 0 |
I mostly run SETI when temperatures get cold and I can use the waste heat to help heat the house. So I'm starting SETI tasks on my computers and am having problems with one machine. Last spring, it was running Lunatics, but when I tried to get tasks, the system said I didn't have a valid SETI app. So I uninstalled Lunatics and installed the original SETI software. I've tried several versions, and they all behave the same. Scheduler requests timeout. Resetting has no apparent effect. I see apps for CUDA 2.3 (6.09), also CUDA_FERMI (6.10) and the stock AstroPulse v6 v6.04 (cuda_opencl_100) version. Since you're using a GT430 (FERMI), the 6.10 (CUDA_FERMI)-app. is the right choice. You've aborted all your work, maybe by RE-Attaching SETI, on this host, (while you're other host with a GT430 works OK), also have Errors on all CUDA 2.3 (6.09) work. Install the same driver used by your other GT430 host and check for older and or driver remnants. A new install of BOINC may also prove usefull, only 1 CUDA_FERMI (6.10) app. should be installed. And 1 for AstroPulse work. TaskManager or ProcessExplorer can be used to see the differences between the 2 GT430 hosts. Your host with the 9600GT/GSO GPUs run probably best on CUDA 2.3 (6.09), IMHO. |
Fred E. Send message Joined: 22 Jul 99 Posts: 768 Credit: 24,140,697 RAC: 0 |
As the kitty man says, the problem is at the server end and we all see what you see. The project's having serious problems with the server that schedules work. Failed work requests that timeout are common now. Here's a few tips: 1) no need to detach/reattach, won't do any good and you abandoned a lot of tasks that will be hard to replace in this environment. 2) Reporting completed work has a higher probability of acceptance if you go on No New Tasks and don't ask for work. After you get them reported, Allow New Tasks. A work request that is not reporting completions seems to have a little better chance of success. 3) Scheduler fills the most capable device first, up to your cache settings. It won't send cpu work while BOINC is asking for gpu work. Lower your cache settings and you may have better luck. There are temporary task limits in place, so a large cache setting doesn't help anyway. You can also turn off gpu work fetch in your Seti@Home prefs on the website. 4) Are you looking for help to fix the Lunatics install? The installer is not available for download now, but if you have it, a reinstall might work. 5) Have patience. They are working on it. Another Fred Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop. |
woodyrox Send message Joined: 7 Apr 01 Posts: 34 Credit: 16,069,169 RAC: 0 |
2) Reporting completed work has a higher probability of acceptance if you go on No New Tasks and don't ask for work. After you get them reported, Allow New Tasks. A work request that is not reporting completions seems to have a little better chance of success. When I turned off new tasks, the update completed instantly. I tried the other tricks and still had no luck getting CPU tasks. Could not find the "gpu fetch" option on the master account preferences page. I can't find my original copy of the Lunatics installer on my network so don't need help trying to install it until the download becomes available. When I last installed it, I thought it gave about 10% better performance. The Lunatics install also made it possible for me to run 6 tasks simultaneously -- 4 CPU and 2 GPU on another machine... the one with the quadcore and the 9600 GSO video card. BTW, that 2x 9600 GSO host really isn't. It's a 9600 GSO and a GeForce 8400 GS. I'm hesitant to mess with the drivers because it's not supposed to work unless they're the same graphic cards. I had a spare 8400 PCI card so I plugged it in and it started crunching along with the 9600. That box only has one PCIe slot so I was happy to get another GPU card running in in. I'll just wait a while and hopefully the server problem gets resolved. None of my other computers were timing out, so I thought it was a problem with that machine. Thanks everyone for your suggestions. |
Fred E. Send message Joined: 22 Jul 99 Posts: 768 Credit: 24,140,697 RAC: 0 |
Could not find the "gpu fetch" option on the master account preferences page. On the website's SETI@Home project prefs, it's "Use Nvidia GPU" (or ATI). That pref doesn't stop crunching what you have. On the next update, Scheduler will tell BOINC to stop asking for gpu work. I usually keep a browser tab open to that page to remind me to reset it when I get the cpu tasks. Good luck - work fetch is tough these days. Another Fred Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop. |
woodyrox Send message Joined: 7 Apr 01 Posts: 34 Credit: 16,069,169 RAC: 0 |
Good news, the updates started working on my faulty box last night. Since then, I've received CPU tasks, and completed tasks are reported very quickly. I'm in need of fixing my Lunatics install on another machine that I'm bringing online. Again, it's complaining about unusable versions. Here's the log: SETI@home, 2012-11-11 10:11:05, Message from server: Your app_info.xml file doesn't have a usable version of SETI@home Enhanced. SETI@home, 2012-11-11 10:11:05, Message from server: Your app_info.xml file doesn't have a usable version of AstroPulse v6. SETI@home, 2012-11-11 10:11:05, Message from server: This computer has finished a daily quota of 1 tasks SETI@home, 2012-11-11 10:16:10, Sending scheduler request: To fetch work. SETI@home, 2012-11-11 10:16:10, Requesting new tasks for CPU SETI@home, 2012-11-11 10:16:18, Scheduler request completed: got 0 new tasks SETI@home, 2012-11-11 10:16:18, Message from server: No tasks sent SETI@home, 2012-11-11 10:16:18, Message from server: Your app_info.xml file doesn't have a usable version of SETI@home Enhanced. SETI@home, 2012-11-11 10:16:18, Message from server: This computer has finished a daily quota of 1 tasks Here is the content of my app_info.xml file: <app_info> <app> <name>setiathome_enhanced</name> </app> <file_info> <name>AK_v8b_win_SSE3_AMD.exe</name> <executable/> </file_info> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>603</version_num> <file_ref> <file_name>AK_v8b_win_SSE3_AMD.exe</file_name> <main_program/> </file_ref> </app_version> <app> <name>astropulse_v505</name> </app> <file_info> <name>ap_5.05r409_SSE.exe</name> <executable/> </file_info> <app_version> <app_name>astropulse_v505</app_name> <version_num>505</version_num> <file_ref> <file_name>ap_5.05r409_SSE.exe</file_name> <main_program/> </file_ref> </app_version> <app> <name>setiathome_enhanced</name> </app> <file_info> <name>Lunatics_x38g_win32_cuda32.exe</name> <executable/> </file_info> <file_info> <name>cudart32_32_16.dll</name> <executable/> </file_info> <file_info> <name>cufft32_32_16.dll</name> <executable/> </file_info> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>610</version_num> <plan_class>cuda_fermi</plan_class> <avg_ncpus>0.040000</avg_ncpus> <max_ncpus>0.040000</max_ncpus> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>Lunatics_x38g_win32_cuda32.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>cudart32_32_16.dll</file_name> </file_ref> <file_ref> <file_name>cufft32_32_16.dll</file_name> </file_ref> </app_version> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>609</version_num> <plan_class>cuda23</plan_class> <avg_ncpus>0.040000</avg_ncpus> <max_ncpus>0.040000</max_ncpus> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>Lunatics_x38g_win32_cuda32.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>cudart32_32_16.dll</file_name> </file_ref> <file_ref> <file_name>cufft32_32_16.dll</file_name> </file_ref> </app_version> <app_version> <app_name>setiathome_enhanced</app_name> <version_num>608</version_num> <plan_class>cuda</plan_class> <avg_ncpus>0.040000</avg_ncpus> <max_ncpus>0.040000</max_ncpus> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>Lunatics_x38g_win32_cuda32.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>cudart32_32_16.dll</file_name> </file_ref> <file_ref> <file_name>cufft32_32_16.dll</file_name> </file_ref> </app_version> </app_info> Any ideas what the problem might be and how to fix it? I could uninstall Lunatics and start over like I did on the other box. This machine has a quad core CPU and only 1 GPU so it would work with a standard SETI install, albeit at slightly lower performance. If possible, I would like to keep the Lunatics install. Many thanks in advance. |
woodyrox Send message Joined: 7 Apr 01 Posts: 34 Credit: 16,069,169 RAC: 0 |
I compared my app_info.xml file on the failing machine with another flawlessly crunching, AMD 9850 quad core and they are identical. All of the applications listed in the two app_info.xml files also exist in the seti projects folder: C:\Documents and Settings\All Users\Application Data\BOINC\projects\setiathome.berkeley.edu So I'm confused why one machine works fine and the other complains about an unusable app version when everything looks identical between the two. Has anyone seen this before? |
woodyrox Send message Joined: 7 Apr 01 Posts: 34 Credit: 16,069,169 RAC: 0 |
OK, the astropulse v6 message is probably because that's the only astropulse work that's currently available. The AMD Lunatics install on the functional box isn't getting any Astropulse work either. But it's not complaining about it. The unusable enhanced app problem I still don't get. Is it possible to edit the app info file to add a v6 astropulse app and manually download the corresponding app? I don't want to change the install on my 6-processor box because the SETI app had the gpu + cpu ratio set such that only 5 processors would run simultaneously. Thanks. |
©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.