First off, you had download errors for the application files needed to process work in the different science runs here on Eintein@home. Something between Boinc on your computer and the project download server prevented you from downloading the necessary files, the most likely reason for this is as others have mentioned either an overaggressive antivirus program or an miss configured firewall, either software or hardware (ie router). It has absolutely nothing to do with app_info.xml.
If you have the files downloaded on another machine you could just copy them over to the affected machine to skip the download step but you would probably run into the same problem next time new versions are released.
You could also try to download the files via a web-browser by pasting something like this into the address bar: http://einstein2.aei.uni-hannover.de/download/einsteinbinary_BRP4G_1.52_windows_intelx86__BRP4G-Beta-cuda32-nv301.exe to see if that works on the affected machine.
Use venues/locations to separate preferences for different machines. Got to your prefs and created separate prefs for either home/work/school as to your liking. Then go to your computers page and click details for the host your wish to assign to the new prefs. At the bottom of the details page there's a location selection box to change the location of the computer.
However, all the downloads apparently failed - (noted in my Tasks Mgr)
then files removed by BOINC via update and "reported" - returning my Cache(s) to Empty
And this persistent Event Log:
2/6/2016 1:09:13 PM | Einstein@Home | Requesting new tasks for CPU and NVIDIA GPU
2/6/2016 1:09:16 PM | Einstein@Home | Scheduler request completed: got 0 new tasks
2/6/2016 1:09:16 PM | Einstein@Home | No work sent
2/6/2016 1:09:16 PM | Einstein@Home | No work is available for Binary Radio Pulsar Search (Arecibo, GPU)
2/6/2016 1:09:16 PM | Einstein@Home | No work is available for Binary Radio Pulsar Search (Parkes PMPS XT)
2/6/2016 1:09:16 PM | Einstein@Home | (reached daily quota of 33 tasks)
2/6/2016 1:09:16 PM | Einstein@Home | Project has no jobs available
At the time you reported the above, did you go to the website and look at the tasks list for one of your computers? For example, this is the list for hostID 12125768. You could then pick any failed task, click on its Task ID link to see the stderr output. Below is the output for one such task chosen at random.
7.6.22
app_version download error: couldn't get input files:
You can see very clearly that there were two files failing to download and that these were both .dll files that are necessary for the new app and its tasks to run. If BOINC is being prevented from downloading these files, then all the tasks of this type assigned to your host will immediately fail with a computation error and then be reported to the server as such at the next scheduler contact.
The only problem you need to solve is why the downloading of these two files was being interfered with.
All the other actions you have taken or have suggested taking are just going to make things worse. If you don't fix what is interfering with the .dll downloads, you won't get BRP4G work. Resetting the project, changing preferences randomly, changing app_config.xml, using a registry cleaner, rebooting the machine, etc, don't seem to be in any way connected to the actual problem - the bit of software that is deciding it doesn't like those .dlls and is blocking the downloads.
When something like this goes wrong, the very worst thing to do is adopt the scatter-gun approach of blasting everything else in sight. Before you do anything in haste, you need to properly identify the issue. You need to learn to use the tools on the website (like stderr.txt or the last scheduler contact as two examples) and if you don't understand the information given, post that information and ask for an explanation.
Your next course of action should be to really search for what it is that is blocking the downloads and fix that.
Hi Folks,
Dunno whats up but since started getting and crunching BRP4G task I've had a slew of invalids, I still have a few WU left, but have removed BRP4G from my settings.
I've been going bonkers trying to work out what the problem is, tried setting my bios back to defaults and no o/c removed 16gig of DDR3 2400 I'd fitted recently in case that was the problem, but it persisted:-(
Repaired BOINC in case that had gone sideways, but I still ended up with 1 WU validating ok, and then a couple of WU later another invalid..
Since BRP6 task are available again I'll stick to those unless the problem persists.
Since I don't know the reason for getting invalids only since the 6th when BRP4G became available I'm hoping its just those affected.
I'd be tempted to retry running BRP4G at x1,and if still generating invalids then look at the driver version. Tricky problem.
Already done that, went back to a former working bios, I suspected it might have been the hotfix Nvidia released recently, but situation failed to improve.
However I've had consecutive BRP6 tasks validate ok, so I'm still without a solution other than to not use BPR4G WU.As for running X1, that's what I usually do, and was doing when I got those invalids.
I also dedicate a GTX980ti to individual projects and I've swapped GPU assigned to the project.. Like I posted its driving me somewhat bonkers:-/
At present I'm working through bios settings to try and get my projects back up to speed, but its a slow process, since there a good few of those settings, including some not so obvious ones that can affect performance / PCIE x16 speed and stability..
Anyway as long as BRP6 tasks continue to validate without error I'll crunch them exclusively and I agree its a darn tricky problem:-(
Well I've found the problem! My Zotec X980ti is borked, it just dropped out of P2 and wont allow it again:-(
I'll have to trace paperwork to see if I cam get a RMA for the dratted thing:-(
In the meantime its crunching MW@H at over 5 mins per WU:-( Usually its 54 seconds..
Well I've found the problem! My Zotec X980ti is borked, it just dropped out of P2 and wont allow it again:-(
I'll have to trace paperwork to see if I cam get a RMA for the dratted thing:-(
In the meantime its crunching MW@H at over 5 mins per WU:-( Usually its 1 min 54 seconds..
Regards,
ARRGGHHH... Took card out, put card back in, and its working again. To top it off the firm I got it from no longer sells that card...
I suspect your hosts are
)
I suspect your hosts are deemed "unreliable" because of the large number or error tasks - the scheduler will restrict the number of tasks per day.
Looking at your hosts
"Last contact" shows for example one host
2016-02-07 21:06:46.8623 [PID=22645] [send] [HOST#12125769] not reliable; max_result_day 2
You need to let it become reliable, and more tasks will be assigned.
First off, you had download
)
First off, you had download errors for the application files needed to process work in the different science runs here on Eintein@home. Something between Boinc on your computer and the project download server prevented you from downloading the necessary files, the most likely reason for this is as others have mentioned either an overaggressive antivirus program or an miss configured firewall, either software or hardware (ie router). It has absolutely nothing to do with app_info.xml.
If you have the files downloaded on another machine you could just copy them over to the affected machine to skip the download step but you would probably run into the same problem next time new versions are released.
You could also try to download the files via a web-browser by pasting something like this into the address bar: http://einstein2.aei.uni-hannover.de/download/einsteinbinary_BRP4G_1.52_windows_intelx86__BRP4G-Beta-cuda32-nv301.exe to see if that works on the affected machine.
Use venues/locations to separate preferences for different machines. Got to your prefs and created separate prefs for either home/work/school as to your liking. Then go to your computers page and click details for the host your wish to assign to the new prefs. At the bottom of the details page there's a location selection box to change the location of the computer.
RE: Yes, thanks Christian -
)
At the time you reported the above, did you go to the website and look at the tasks list for one of your computers? For example, this is the list for hostID 12125768. You could then pick any failed task, click on its Task ID link to see the stderr output. Below is the output for one such task chosen at random.
app_version download error: couldn't get input files:
cudart_xp32_32_16.dll
-224 (permanent HTTP error)
permanent HTTP error
cufft_xp32_32_16.dll
-224 (permanent HTTP error)
permanent HTTP error
]]>
You can see very clearly that there were two files failing to download and that these were both .dll files that are necessary for the new app and its tasks to run. If BOINC is being prevented from downloading these files, then all the tasks of this type assigned to your host will immediately fail with a computation error and then be reported to the server as such at the next scheduler contact.
The only problem you need to solve is why the downloading of these two files was being interfered with.
All the other actions you have taken or have suggested taking are just going to make things worse. If you don't fix what is interfering with the .dll downloads, you won't get BRP4G work. Resetting the project, changing preferences randomly, changing app_config.xml, using a registry cleaner, rebooting the machine, etc, don't seem to be in any way connected to the actual problem - the bit of software that is deciding it doesn't like those .dlls and is blocking the downloads.
When something like this goes wrong, the very worst thing to do is adopt the scatter-gun approach of blasting everything else in sight. Before you do anything in haste, you need to properly identify the issue. You need to learn to use the tools on the website (like stderr.txt or the last scheduler contact as two examples) and if you don't understand the information given, post that information and ask for an explanation.
Your next course of action should be to really search for what it is that is blocking the downloads and fix that.
Cheers,
Gary.
BRP6 tasks are flowing again!
)
BRP6 tasks are flowing again!
Cheers,
Gary.
Hi Folks, Dunno whats up but
)
Hi Folks,
Dunno whats up but since started getting and crunching BRP4G task I've had a slew of invalids, I still have a few WU left, but have removed BRP4G from my settings.
I've been going bonkers trying to work out what the problem is, tried setting my bios back to defaults and no o/c removed 16gig of DDR3 2400 I'd fitted recently in case that was the problem, but it persisted:-(
Repaired BOINC in case that had gone sideways, but I still ended up with 1 WU validating ok, and then a couple of WU later another invalid..
Since BRP6 task are available again I'll stick to those unless the problem persists.
Since I don't know the reason for getting invalids only since the 6th when BRP4G became available I'm hoping its just those affected.
Apologies to those wingpersons involed.
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
RE: Since I don't know the
)
I'd be tempted to retry running BRP4G at x1,and if still generating invalids then look at the driver version. Tricky problem.
Please suggest an app_config
)
Please suggest an app_config entry to restrict intel_gpu (usage)to >1 My current prefs are set to .33 But that applies to my discrete GPUs
My intel_gpu is not nearly as strong 8 hours runtime estimate is incompatible
and I don't know what CPU usage is prescribed by Dev either
Please help
@AgentB RE: I'd be
)
@AgentB
Already done that, went back to a former working bios, I suspected it might have been the hotfix Nvidia released recently, but situation failed to improve.
However I've had consecutive BRP6 tasks validate ok, so I'm still without a solution other than to not use BPR4G WU.As for running X1, that's what I usually do, and was doing when I got those invalids.
I also dedicate a GTX980ti to individual projects and I've swapped GPU assigned to the project.. Like I posted its driving me somewhat bonkers:-/
At present I'm working through bios settings to try and get my projects back up to speed, but its a slow process, since there a good few of those settings, including some not so obvious ones that can affect performance / PCIE x16 speed and stability..
Anyway as long as BRP6 tasks continue to validate without error I'll crunch them exclusively and I agree its a darn tricky problem:-(
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
@AgentB Well I've found
)
@AgentB
Well I've found the problem! My Zotec X980ti is borked, it just dropped out of P2 and wont allow it again:-(
I'll have to trace paperwork to see if I cam get a RMA for the dratted thing:-(
In the meantime its crunching MW@H at over 5 mins per WU:-( Usually its 54 seconds..
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
RE: @AgentB Well I've
)
ARRGGHHH... Took card out, put card back in, and its working again. To top it off the firm I got it from no longer sells that card...
blasted computers... and Murphy:-(
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
