System with one RTX A4500 and dual A6000 (driver: 516.94): 50 received, every single work unit failed (1015 (0x000003F7) Unknown error code).
System with dual Quadro RTX 6000 (driver: 516.94): 25 received, 5 completed and validated, the rest are computation errors.
System with a single RTX A4500 (driver: 516.94): 38 received, 10 completed and validated, the rest are computation errors.
It is odd that the first system in the list didn't get through any of the 50 work units yet the other systems did. Or, it might not be odd... really not sure. Same GPU drivers across these systems.
System with one RTX A4500 and dual A6000 (driver: 516.94): 50 received, every single work unit failed (1015 (0x000003F7) Unknown error code).
System with dual Quadro RTX 6000 (driver: 516.94): 25 received, 5 completed and validated, the rest are computation errors.
System with a single RTX A4500 (driver: 516.94): 38 received, 10 completed and validated, the rest are computation errors.
It is odd that the first system in the list didn't get through any of the 50 work units yet the other systems did. Or, it might not be odd... really not sure. Same GPU drivers across these systems.
I wouldn't worry about it too much, likely nothing wrong with your system. over the past week, the project has sent out many different versions of the applications for BRP7 for both Windows and Linux. some worked and some didn't. so depending which application you processed with at which times, would determine if they were successful or not. probably just bad luck that the first machine never got a successful run, or that system was busy working on something else to request work from BRP7 when the successful apps were being distributed.
these are still in flux for beta testing. expect to have varying levels of success with the application(s) until they hammer out the finer details to get a stable app.
I issued new workunits with slightly changed command-line parameters the OpenCL apps should be able to cope with. I enabled the OpenCL app version for Windows (AMD and NVidia) and Linux AMD. These should now only get the new tasks. I also enabled Linux-CUDA, as this is the only version that should be able to finish the remaining older workunits.
I notice the server status page shows BRP7 task "in progress" remains very low at 150.
My Windows/AMD host which received two BRP7 tasks about an hour ago has been getting denied more with indications like this:
2022-08-26 12:49:06.9781 [PID=15991] [send] [HOST#12260865] will accept beta work. Scanning for beta work.
2022-08-26 12:49:07.0105 [PID=15991] [send] [HOST#12260865] beta work found: [RESULT#1341574387]
2022-08-26 12:49:07.0106 [PID=15991] [version] Checking plan class 'BRP7-opencl-ati'
2022-08-26 12:49:07.0122 [PID=15991] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
2022-08-26 12:49:07.0122 [PID=15991] [version] WU#666190219 too old
2022-08-26 12:49:07.0122 [PID=15991] [version] Checking plan class 'BRP7-opencl-nvidia'
2022-08-26 12:49:07.0122 [PID=15991] [version] WU#666190219 too old
2022-08-26 12:49:07.0123 [PID=15991] [version] no app version available: APP#57 (einsteinbinary_BRP7) PLATFORM#9 (windows_x86_64) min_version 0
2022-08-26 12:49:07.0123 [PID=15991] [version] no app version available: APP#57 (einsteinbinary_BRP7) PLATFORM#2 (windows_intelx86) min_version 0
...
2022-08-26 12:49:07.0604 [PID=15991] [debug] [HOST#12260865] MSG(high) No work sent
2022-08-26 12:49:07.0604 [PID=15991] [debug] [HOST#12260865] MSG(high) No work is available for Binary Radio Pulsar Search (MeerKAT)
Ian&Steve C. wrote: try this
)
Thanks, that works perfectly. I was using the link from within Boinc for "your results", which should perhaps be updated? it goes to https://einsteinathome.org/account/tasks/active I've created this thread in bugs: https://einsteinathome.org/content/outdated-link-results
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
I have run the 1.70 (beta)
)
I have run the 1.70 (beta) Opencl-intel_gpu-newer app and have 37 pending and 1 invalid that did not match the 1.33 app result. No valid results yet.
Currently 17 invalid and no valid.
I am a little late on this
)
I am a little late on this but our results:
System with one RTX A4500 and dual A6000 (driver: 516.94): 50 received, every single work unit failed (1015 (0x000003F7) Unknown error code).
System with dual Quadro RTX 6000 (driver: 516.94): 25 received, 5 completed and validated, the rest are computation errors.
System with a single RTX A4500 (driver: 516.94): 38 received, 10 completed and validated, the rest are computation errors.
It is odd that the first system in the list didn't get through any of the 50 work units yet the other systems did. Or, it might not be odd... really not sure. Same GPU drivers across these systems.
Boca Raton Community HS
)
I wouldn't worry about it too much, likely nothing wrong with your system. over the past week, the project has sent out many different versions of the applications for BRP7 for both Windows and Linux. some worked and some didn't. so depending which application you processed with at which times, would determine if they were successful or not. probably just bad luck that the first machine never got a successful run, or that system was busy working on something else to request work from BRP7 when the successful apps were being distributed.
these are still in flux for beta testing. expect to have varying levels of success with the application(s) until they hammer out the finer details to get a stable app.
_________________________________________________________________________
Is a automatic timeout (2h or
)
Is a automatic timeout (2h or so? ) possible for the WUs? My GTX 1650 run a WU for 32h with an error at the end.
https://einsteinathome.org/de/task/1341574599
I issued new workunits with
)
I issued new workunits with slightly changed command-line parameters the OpenCL apps should be able to cope with. I enabled the OpenCL app version for Windows (AMD and NVidia) and Linux AMD. These should now only get the new tasks. I also enabled Linux-CUDA, as this is the only version that should be able to finish the remaining older workunits.
BM
IAN&STEVE C.- thanks for the
)
IAN&STEVE C.- thanks for the insight, we really appreciate it.
BERND- thanks for the update and we will see how they go when we receive some of them.
I notice the server status
)
I notice the server status page shows BRP7 task "in progress" remains very low at 150.
My Windows/AMD host which received two BRP7 tasks about an hour ago has been getting denied more with indications like this:
2022-08-26 12:49:06.9781 [PID=15991] [send] [HOST#12260865] will accept beta work. Scanning for beta work.
2022-08-26 12:49:07.0105 [PID=15991] [send] [HOST#12260865] beta work found: [RESULT#1341574387]
2022-08-26 12:49:07.0106 [PID=15991] [version] Checking plan class 'BRP7-opencl-ati'
2022-08-26 12:49:07.0122 [PID=15991] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
2022-08-26 12:49:07.0122 [PID=15991] [version] WU#666190219 too old
2022-08-26 12:49:07.0122 [PID=15991] [version] Checking plan class 'BRP7-opencl-nvidia'
2022-08-26 12:49:07.0122 [PID=15991] [version] WU#666190219 too old
2022-08-26 12:49:07.0123 [PID=15991] [version] no app version available: APP#57 (einsteinbinary_BRP7) PLATFORM#9 (windows_x86_64) min_version 0
2022-08-26 12:49:07.0123 [PID=15991] [version] no app version available: APP#57 (einsteinbinary_BRP7) PLATFORM#2 (windows_intelx86) min_version 0
...
2022-08-26 12:49:07.0604 [PID=15991] [debug] [HOST#12260865] MSG(high) No work sent
2022-08-26 12:49:07.0604 [PID=15991] [debug] [HOST#12260865] MSG(high) No work is available for Binary Radio Pulsar Search (MeerKAT)
Yep, too many "old" WUs in
)
Yep, too many "old" WUs in the queue. Should get better with time when these are finished more and more.
BM
I broke up a WU that run
)
I broke up a WU that run 8,5h....so pay attention, that your systems running not to long on this kind of WUs.
https://einsteinathome.org/de/task/1341565200