Anyone know whats going on with the credits for Gravitational Wave Search for the GPU's?
There's nothing "going on" with GW task credits. Everything is normal. Your problem is that you're not doing very many tasks.
Take a look at your tasks list. Your entire list shows 79 valid tasks for several weeks. In the last 24 hours there were 5 so a rough estimate, if that rate of production were to continue, is that your RAC will eventually get down to around 5K. You could easily work this out for yourself.
If you want a higher Einstein RAC, run more tasks. Your equipment is certainly capable of doing so - two GTX 1080s.
Ok..so I must have seen a spike when I came back online after a CPU failure that took me offline for a few weeks.
The system as you might notice is used on many projects, so I am maxed out. I have many interests so I have many projects. After further examination of the goals of all the projects, I am dropping one of the GPU projects as there is no scientific value in its work.
The system as you might notice is used on many projects, ...
Yes, of course others can see that you have quite a list :-).
But what is not easy for others to know is what has been happening to your RAC at these other projects. Unless your machine has a new problem, the drop in Einstein RAC should have been offset by some sort of rise elsewhere.
On my hosts, there are 'job_log' files in the BOINC directory for projects the host contributes to. There is a dated entry for each returned task so it would be possible to tally how many tasks are returned for a particular time period for each different project that runs. The main problem is that the date is in 'Unix epoch' format so you would need to convert these into human readable dates. On-line converters can do that.
If you picked a date - say two weeks ago - you could just total the entries for each job_log file (from then to now) to see the split of numbers for the different projects you are using. You could then see if some project was getting more or less than its expected 'share' of your system. Changes in RAC would then be easy to explain.
Gary Roberts wrote:Greg_BE
)
Ok..so I must have seen a spike when I came back online after a CPU failure that took me offline for a few weeks.
The system as you might notice is used on many projects, so I am maxed out. I have many interests so I have many projects. After further examination of the goals of all the projects, I am dropping one of the GPU projects as there is no scientific value in its work.
Thanks for the explanation about the credits.
Greg_BE wrote:The system as
)
Yes, of course others can see that you have quite a list :-).
But what is not easy for others to know is what has been happening to your RAC at these other projects. Unless your machine has a new problem, the drop in Einstein RAC should have been offset by some sort of rise elsewhere.
On my hosts, there are 'job_log' files in the BOINC directory for projects the host contributes to. There is a dated entry for each returned task so it would be possible to tally how many tasks are returned for a particular time period for each different project that runs. The main problem is that the date is in 'Unix epoch' format so you would need to convert these into human readable dates. On-line converters can do that.
If you picked a date - say two weeks ago - you could just total the entries for each job_log file (from then to now) to see the split of numbers for the different projects you are using. You could then see if some project was getting more or less than its expected 'share' of your system. Changes in RAC would then be easy to explain.
Cheers,
Gary.