FGRPB1G - No tasks to send

Hi Martin,

If you don't give any information, it's a lot harder to think through all the possible reasons.  I don't see any particular problem with the project issuing tasks, so the short answer is that it must be something to do with the way you have your machine configured.  Here are some things to consider.

1. What particular searches have you selected to run?
2. Have you previously been getting tasks for ALL selected searches?
3. Have you made any preference settings changes recently?  If so, what?
4. As you have been running FGRPB1 CPU tasks, are you aware of the upcoming transition to FGRP5 for CPU work?
5. Have you previously been getting gravity wave (O1...) tasks?  Is that search still selected in your prefs?
6. Have you perhaps run out of disk space BOINC is allowed to use?
7. Have you checked the scheduler logs to see if any reasons are given there?

I provided a couple of links to information that might be relevant to your question.  I suggest you read them both carefully to see if they might solve your problem.

I decided to check the scheduler log for your machine's last contact with the server and it looks like you have too severe a restriction on disk space.  Here are some relevant lines from the log.  It looks like the scheduler wants to send you some GW work but you don't have enough disk space that BOINC is allowed to use.  In the excerpt below, I've trimmed the date, time and PID information so that the actual messages will be more readable (lines don't overflow).  I've added some highlighting to the important bits.

    [send] CPU: req 43560.00 sec, 2.00 instances; est delay 0.00
    [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
    [send] Intel GPU: req 0.00 sec, 0.00 instances; est delay 0.00
    [send] work_req_seconds: 43560.00 secs
    [send] available disk 0.09 GB, work_buf_min 0
    [send] active_frac 0.998437 on_frac 0.998212 DCF 0.841335
    [mixed] sending locality work first (0.4432)
    [send] send_old_work() no feasible result older than 336.0 hours
    [version] Checking plan class 'AVX'
    [version] reading plan classes file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
    [version] plan class ok
    [version] Best version of app einstein_O1Spot1Lo is 1.00 ID 986 AVX (7.52 GFLOPS)
    [send] send_old_work() no result younger than 263.3 hours and older than 168.0 hours
    [send] stopping work search - insufficient disk space

 As a first step, I suggest you go to your account page and select the 'preferences' menu item.  This will bring you to computing preferences.  If you are using different 'locations' (venues) make sure you have the right one selected.  You need to open the 'Advanced Settings' drop down link and look for Disk and Memory settings and tell us what you have the three disk related values set to.  These are the settings referred to in the link I provided earlier in this reply.

If you need help with adjusting these, please ask.

 If you have previously done a lot of GW work, the accumulation of large data files for that work might be the cause of your current lack of available space.  If you can't afford to allow BOINC to use more disk space, you can't just reclaim space by deleting what you might think are 'old and no longer needed files' since the scheduler will be likely to just send them all back to you.  In that case, you should just reset the project since you have no 'in progress' work at the moment and that will give you a fresh start with just the files you need for the future.

There shouldn't be any problems getting FGRPB1 CPU tasks. The new search is still in testing and shouldn't impact the other searches. We didn't exclude CPUs from the FGRPB1 search intentionally. If you don't get tasks please open a new thread and provide the hostids of computers where you think you should get FGRPB1 CPU work. The drop of in progress tasks has other reasons which are understood and do not point to a problem.