Script it by creating multiple global_prefs_override.xml files then set scheduler/crontab jobs to copy and read them in with boinccmd.

Do we want these features of day/nighttime Memory, CPU cores numbers used and throttles? No doubt about it.

--//--

or could you add this feature to the next release ?

The devs will look into it for a future version, just not the next one. See their Remodel of computing preferences for more information.

Thanks

Hi,

I have a couple of minor wishes, most preferred first:

1) Set a "Run For..." time. For example, I'd like to run at night, but only for a certain number of hours when I'm asleep (so i don't run it all night and overtax my computer/to give it time to sleep). I'd like to give it a command that says "Run For X Hrs" then stop, instead of setting a hard daily time (as schedules vary..)

2) Upgraded menu bar item (for the mac). Add features like "Suspend/Run" (not just snooze), "Suspend For...," "Run For...,"

3) Option to run application Only in dock or Only in menu bar item (Mac). I've found suggestions for not having any user interface at all (run as system process), but I would just like to have one interface, as having both seems repetitive. Possibly add an option to have it run only with a menu bar item and give an option to open full manager app window.

4) Track total data transfer over a certain time. I don't want to set a transfer limit unless I have to.

When a user selects No New Tasks for a project, it would be useful to have these tasks automatically report when they complete.
All complete or individually? Either would be better than the present situation; 6.10.60 does not automatically report tasks even when all tasks have finished and uploaded.

6.10.60 does not automatically report tasks even when all tasks have finished and uploaded.

Have you waited 24 hours?

Completed work is reported at the first of:

1) 24 hours before deadline
2) Connect Every X before deadline.
3) 24 hours after task completion.
4) Immediately if the upload completes later than either 1, 2, or 3 upon completion of the task.
5) On a trickle up message (CPDN only, I believe).
6) On a trickle down request.
7) On a server scheduled connection. Used, but I am not certain by which project.
8) On a request for new work.
9) When the user pushes the update button.
10) On a request from an account manager.

I'm reporting a minor default functionality problem here. This problem has popped up in several threads recently, when the average user discovers their tasks finished but did not report.

If for example, someone selects no new tasks because they are going on holidays for 3weeks and then just shuts the computer down when existing tasks have finished (100%) and uploaded, the tasks will still report too late by the time they get back from holiday (for most projects).

So I am asking that a new default setting be created to automatically report tasks when the last task is completed, in the situation where the cruncher chooses No new tasks.

For long tasks (over 12h) it might be a good idea that these report immediately, but that's another issue. Ideally each project could be individually configured to report immediately or not.

I am aware of the cc_config.xml file and how to use report tasks immediately. This setting is recommended for some GPU projects but not welcomed by some CPU projects.

---

From client configuration, there's a command-line option

--exit_when_idle
Exit when there are no more tasks, and report completed tasks immediately. New in 6.1

which is close to what you want, though not as 'user friendly' (!) as editing cc_config.xml

Thanks Richard,

That would be perfect if it could be introduced into the Boinc Manager; an exit_when_idle button/option (Close Boinc When Tasks Finish), ideally with a text tip advising that this reports all tasks and closes Boinc.
Possibly with an additional, and shut down windows option.

I have 1 suggestion:


To allow tasks flagged as non-cpu-intensive to continue running under different performance regulations(Like to run when computer is in use or have a different CPU/memory threshold, while regular projects don't do that)

What about rethink BOINC so that the entire subsystem runs as a virtual machine?
I mean something like the new LHC@home 2.0, but instead of have BOINC running a virtual machine, BOINC itself should be a virtual machine where to run different projects.
I think that in this manner many problems related to the use of different OS/different drivers can be avoided or, at least, project developers can easily test out problems by having a standard machine as background.
What do you think about this?

What about rethink BOINC so that the entire subsystem runs as a virtual machine?

Other than plenty of other problems, here's one that will be shot down by plenty of people out there already:
You won't be able to run any work on GPUs or (other) co-processors in the VM. The VM program emulates what hardware to use, which in the case of VirtualBox and the videocard is an S3 Trio32/64 with a maximum of 128MB RAM. Then it doesn't matter if you have one or more GTX570 GPUs in your system, the VM won't use it/them to do science with.

The VM program can use the GPU, but that's probably not what people mean. Then the VM would run on the GPU(s), not run the science programs in the VM on the GPU(s).

Using a VM is still in its infancy in BOINC. There's no real support for it yet in the present recommended release versions. Testing is underway in the latest alpha, but the one project running with VMs hasn't updated to the latest thrills and frills yet, so testing of the client goes slow.

Still, plans are afoot to integrate a VM application such as Virtual Box into BOINC. This is something for the future though. Will be continued.

On the CUDA and switching users problem: Does the application receive a sufficiently precise error code that it can tell when this happened?

If so, BOINC should suspend the spplicatio and displayh a warning. Otherwise, every pending GPU task will error off.

++PLS

---

I don't know if this is possible with the way BOINC is set up, but the things that would be cool to see:
1) Prioritize tasks based on deadline...preemptively increase priority on tasks that with a bit of a time buffer after examining deadline on other tasks.

For instance: Task #1 will take 2d to complete and has a deadline of 3d. Task #2 will take 3d to complete and has a deadline of 8d...so the client should have task #1 run non-stop until completion and then run task #2. An exception could be if task #1 will definitely not complete on time, but task #2 has a chance to or will complete if it runs immediately, then the client could just give up on task #1 and focus on task #2. This type of calculation can be fine-tuned somewhat by monitoring stats over a period of time on each project run to see how the computer behaves on average.

2) If the queue is getting full and it has almost reached a maximum work capacity for the time alloted, even after reporting work units, the client should throttle getting new tasks until the work queue has emptied out some. This also could be fine-tuned by past history.

3) Take into account total time in the calculation of maximum work possible by factoring in competition for GPU resources as well as work units that are set to use multiple CPUs (like some of the Milkyway units which gobble up all of them). Additionally, "high priority" work units should not always run if it is more efficient to run other work units that are due earlier. (Primegrid often locks processors so other work units won't run and might pass its deadline even though there may be plenty of time to run them afterwards.)

In any case, a lot of the projects don't play nice with each other, and I find myself having to babysit them manually which is lame. Just some ideas. Sorry my thought processes aren't completely clear and coalesced. Figured I'd get the general gist out and let the more software engineery/mathematician types see if it's feasible, desirable, and if so, the best way to implement them. I guess really what I'm asking is to have a best fit algorithm added. Or maybe Tetris. :)

Hey clipfox, if you've read carefully, I've navigated you to the "BOINC Dev mailing list" URL, not here to "BOINC Dev message boards" forum, I hope you understand the difference.

Posting your ideas here you go the risk of not presenting them to BOINC developers at all - they take a look here nearly as rarely as once in a blue moon... but they check their e-mails continuously.

Peter

Ah, I just went back to the other forums. I didn't see the reply until now. I'll go repost this at the link you sent me.

Besides of simply rotating thru the projects by time-slices I would like to have the following:

An option to set a project to "fetch work and run ony if active projects are out of work packets". That would give the option to run my preferred project with max priority without going idle if that project is down.

Already available.
Set the favorite project's resource share to anything you want, set the backup project's resource share to zero. This way the backup project will only fetch work when the favorite project does not have any work.

This does require that the backup project have an up-to-date BOINC back-end and that you are using at least BOINC 6.10.56

"fetch work and run ony if active projects are out of work packets".

The method Ageless describes will fetch work from you backup project only if your preferred project has no work, however you should be warned that if your receive tasks from the backup and receive tasks from the preferred later, tasks from the backup will not be auto-aborted or suspended. (..)

That would not be an adored behaviour, but is acceptable. Maybe that part could be enhanced.