// DBOINCP-300: added node comment count condition in order to get Preview working ?>
Peter Crickman
Joined: 14 Jun 23
Posts: 1
Credit: 1501765
RAC: 467
7 Oct 2023 3:56:57 UTC
Topic 230166
(moderation:
)
I have not received any work for the past three days. Can you explain the contents of the log file that is reported in the BOINC log? https://einsteinathome.org/host/13144288/log
Your link points to the most recent scheduler contact that your host has made. Here are some excerpts that show what the scheduler reports.
[send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00Your host requested 0.00 secs of CPU work. If you want to run any CPU tasks, you need to select a CPU search. Are you trying to run GPU tasks only?
[send] Intel GPU: req 30240.00 sec, 0.00 instances; est delay 0.00Your host requested work for an Intel GPU but there was no GPU work available for your particular type of Intel GPU. I guess you may have been running the gamma-ray pulsar search (FGRPB1G) but that search has now finished. Check the announcement in Tech News.
[version] Checking plan class 'BRP7-opencl-intel_gpu'BRP7 is a search that has essentially replaced FGRPB1G. If you look at the lines that follow this, you can see that your GPU is excluded because it doesn't match particular types of Intel GPU.
There is also the following:-
[version] Checking plan class 'opencl-intel_gpu-new'This is a new plan class. If you follow the log for several lines you will see that the scheduler finds that your Intel GPU doesn't meet the requirements since it has a driver version that exceeds the max allowed.
There is also an even newer plan class in:-
[version] Checking plan class 'opencl-intel_gpu-newer' and the only thing that stops it from sending tasks is that the app for it is in beta test, which you need to allow in your prefs, as shown below:-
[version] beta test app versions not allowed in project prefs.
Please be aware that test apps may crash and burn so don't allow a large work cache size until you see that the app is working correctly. I'm not running any Intel GPUs so I don't know how it performs.
Your link points to the most
)
Your link points to the most recent scheduler contact that your host has made. Here are some excerpts that show what the scheduler reports.
[send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
Your host requested 0.00 secs of CPU work. If you want to run any CPU tasks, you need to select a CPU search. Are you trying to run GPU tasks only?[send] Intel GPU: req 30240.00 sec, 0.00 instances; est delay 0.00
Your host requested work for an Intel GPU but there was no GPU work available for your particular type of Intel GPU. I guess you may have been running the gamma-ray pulsar search (FGRPB1G) but that search has now finished. Check the announcement in Tech News.[version] Checking plan class 'BRP7-opencl-intel_gpu'
BRP7 is a search that has essentially replaced FGRPB1G. If you look at the lines that follow this, you can see that your GPU is excluded because it doesn't match particular types of Intel GPU.There is also the following:-
[version] Checking plan class 'opencl-intel_gpu-new'
This is a new plan class. If you follow the log for several lines you will see that the scheduler finds that your Intel GPU doesn't meet the requirements since it has a driver version that exceeds the max allowed.There is also an even newer plan class in:-
[version] Checking plan class 'opencl-intel_gpu-newer'
and the only thing that stops it from sending tasks is that the app for it is in beta test, which you need to allow in your prefs, as shown below:-[version] beta test app versions not allowed in project prefs.
Please be aware that test apps may crash and burn so don't allow a large work cache size until you see that the app is working correctly. I'm not running any Intel GPUs so I don't know how it performs.
Cheers,
Gary.