BRP4G no WU's

Paul, 

Since your computers are hidden, it's hard to tell much about your computers other than that 1st line.

The second line is referencing Gravity Wave search which ended some time ago.

The files listed in lines 3 and 4 are related to said Gravity Wave.

Which make makes me wonder, when was the last time before now that you ran that particular computer if it was still hanging onto such old work units?

Even though some of your questions have been answered (both in this thread and elsewhere), I'll add some additional information.

Please be careful with the distinction between the different science runs.  There really are three separate searches involving radio pulsar data:-

1. BRP4 - small tasks built from Arecibo data and designed to run on small (essentially mobile) devices, etc.

2. BRP4G - much larger (bundled) tasks (again Arecibo data) designed to run on more powerful discrete GPUs.

3. BRP6 - bundled tasks using the Parkes Multibeam Pulsar Survey (PMPS) data - also only for more powerful GPUs.

For the last 6 months, the 3rd run is what you would have been doing if work was continuously available.  As Bernd has announced, BRP6 has now finished and there will only be a very small number of 'resends' in future - replacements for existing tasks which fail - until all quorums are completed.  Bernd's advice was for people wanting to continue with work for discrete GPUs, to transition to the 2nd option above for the time being.  Please realise that the available data for BRP4G is likely to be limited and now that BRP6 is gone, there are likely to be shortages in work until there is a longer term replacement run for discrete GPUs.  There are recent comments about future GPU runs in the long running 'featured' thread at the top of the Science forum.

The first line in the log snippet you posted refers to 'Arecibo' and not 'Arecibo GPU' so it seems to me that you may be trying to get tasks for a type of device you don't have.  Perhaps you need to make sure you have set your preferences for Arecibo, GPU and not just the plain Arecibo.

As you are now aware, all the other messages are to do with the gravity wave (CPU only) search which was completed some time ago.  That search used locality scheduling.  You are sent lots of large data files which may be used (and re-used) for many tasks over a period of time.  After a run finishes distributing tasks that use this data, it can take many weeks for every single quorum to be finally completed and verified.  It's not surprising that deleting the data files is left until well after the run appears to have ended.  It's pure coincidence that this final cleanup of data (NOT tasks) happens to have occurred just now.  It's totally unrelated to any BRP runs.  It's very pleasing to see this cleanup since quite a lot of disk space is now being reclaimed.

Very recently, the supply of BRP4G tasks has dried up, probably quite temporarily.  There would be lots of GPUs formerly doing BRP6 now putting pressure on the available supply.  Maybe the data available for splitting into workunits was used up, or perhaps the workunit generator stopped for some reason.  This is likely to be fixed when staff return to work on Monday morning, Hannover time.  It certainly doesn't signify the end of BRP4G.

Thanks for that information.  I thought it was rather unusual that no staff had responded so far.

Hm, another day has passed and still no answer from the staff..