1) Message boards : ATLAS application : ATLAS disk space error; set BOINC drive? (Message 42236)
Posted 18 Apr 2020 by doug
Post:
Hi,

I got this error after my first ATLAS task had run for a number of hours:

4/18/2020 2:53:46 PM | LHC@home | Aborting task QJ3MDmuFpjwnsSi4apGgGQJmABFKDmABFKDmEpOYDmABFKDmd46DYn_0: exceeded disk limit: 10392.30MB > 9536.74MB

I have no idea where this "disk limit" is set or where it comes from. I can't find any such number in any of my project preferences, BOINC preference or local Manager preferences. Also, AFAICT, BOINC didn't run out of disk space, as I still have 90G free on that drive. Can anyone explain what's causing this issue and how I can fix it?

I would also like to know if it is possible to change the drive BOINC projects use for their work, preferably per project. The original 5 projects I had when I started with BOINC some years ago used a reasonable amount of disk space that I didn't begrudge. But one of those projects (SETI) is ending its BOINC work, and I decided to add Rosetta and LHC to take its place. However, those 2 new projects more than doubled the amount of disk space being used, so now I'd like to move at least some of the projects to another (physical) drive, which, though slower (it's a traditional disk drive), has much more free disk space. Is there any way to do this? If it would be better for me to ask this in another forum, I will do so.

Thanks.

Doug
2) Message boards : Theory Application : Theory task true completion time? (Message 42233)
Posted 18 Apr 2020 by doug
Post:
Hi Harri,

Thanks for the info. Naturally, less than an hour after posting my question, the task did finish (in about 8.5 hours), and my Task page shows "Completed and validated".

Thanks again.

Doug
3) Message boards : ATLAS application : typical length of time an ATLAS task takes? (Message 42228)
Posted 18 Apr 2020 by doug
Post:
Hi,

I've recently joined lhc@home and gotten my first ATLAS tasks; one of those tasks has now been running for a little over 2 hours. I have also started a Theory task and the BOINC Manager "Remaining (estimated)" time changed from close to 3 hours before it started to now just under 4 days, after already running for well over 8 hours.

Can anyone give me an idea of what the average time to complete an ATLAS task would be? BOINC Manager is saying something over 4 hours with 4 CPUs - is that reliable? I'd rather not have all CPUs dedicated to one task, so is it possible to accurately estimate the time to complete for 1, 2 and 3 CPUs as well?

My machine is latest Windows 10, Intel Core i5-3470 with 16GB memory, 4 cores. I have the latest BOINC and Virtual Box combo installed (BOINC 7.16.5), which I upgraded to a week or so ago.

Thanks.

Doug
4) Message boards : ATLAS application : set ATLAS "# of cpus/jobs" in app_config.xml? (Message 42227)
Posted 18 Apr 2020 by doug
Post:
Hi,

I've recently joined lhc@home and just got my first ATLAS tasks last night (thanks to help on the Number Crunching forum).

I noticed that all of the ATLAS tasks say "Ready to start (4 CPUs)". I assume that means each ATLAS task will use all 4 cores to run. I had that issue with cosmology@home - I don't want all the cores being monopolized by 1 Task/project. I was able to fix that for cosmology@home by creating an app_config.xml and setting "<max_concurrent>" and "<avg_ncpus>". Is it possible to do the same thing for the ATLAS app in lhc@home? If so, what would the exact "<app_name>" be?

I realize I can accomplish most (or all) of this by setting the "Max # CPUs" and "Max # jobs" options in my project preferences, but I was wondering if it could be done (perhaps more granularly) in the app_config.xml.

Thanks.

Doug
5) Message boards : Theory Application : Theory task true completion time? (Message 42226)
Posted 18 Apr 2020 by doug
Post:
Hi,

I've recently joined lhc@home and gotten my first Theory tasks.

I'm currently running one of these Tasks. The BOINC Manager "Remaining (estimated)" time for each of these tasks is very close to 3 hours. However, the one that is running has been going for almost 8 hours and the remaining estimated time is now almost 4 days. Does that seem right? I've seen the BOINC remaining times be off by quite a bit before, but not by multiple days.

Thanks.

Doug[/b]
6) Message boards : Number crunching : Not getting any tasks, though many are available (Message 42220)
Posted 18 Apr 2020 by doug
Post:
Hi all,

I suspended all my other BOINC projects and Updated lhc@home, and immediately got 16 new tasks, 8 ATLAS and 8 Theory. I've got 1 Theory task running now, and so far, so good.

I do notice that all of the ATLAS tasks say "Ready to start (4 CPUs)". I assume that means (as I remember it did with cosmology@home) that it will use all 4 cores to run 1 task. I had that issue with cosmology@home - I don't want all the cores being monopolized by 1 Task/project. I was able to fix that for cosmology by creating an app_config.xml and setting "<max_concurrent>" and "<avg_ncpus>". Is it possible to do the same thing for the ATLAS app in lhc@home? If so, what would the exact "<app_name>" be?

I also agree with the other posters here who suggested adding additional help/info to the LHC preferences to explain what the "Run native if available" option is and what to set it to for various OSs, especially if it is going to be set "true" by default (which it seems to be - I don't remember setting it).

Thanks again to all here for all the help.

Doug
7) Message boards : Number crunching : Not getting any tasks, though many are available (Message 42219)
Posted 18 Apr 2020 by doug
Post:
Done. I'll wait and see if I get any more tasks.

Thanks.

Doug
8) Message boards : Number crunching : Not getting any tasks, though many are available (Message 42215)
Posted 17 Apr 2020 by doug
Post:
Hi computerzrmle,

I do have "Run native if available" checked. Can you explain how that is relevant to me not getting work?

Thanks.

Doug
9) Message boards : Number crunching : Not getting any tasks, though many are available (Message 42210)
Posted 17 Apr 2020 by doug
Post:
Hi maeax,

My apologies. I didn't fully specify my setup. Also, I just set my computer to show on the site.

I have the latest BOINC and Virtual Box combo installed (BOINC 7.16.5), as of a week or so ago when that upgrade became available on the BOINC site. I've run another Virtual Box project, cosmology@home, for years. I don't think that is my problem.

As of right now, the Server Status page is showing 1311 Theory tasks and 2418 ATLAS tasks as "unsent". I'm still wondering wny I'm not getting any of those. The BOINC Event Log just shows:

"No tasks sent
Scheduler request completed: got 0 new tasks
Project requested delay of 6 seconds"

Thanks.

Doug
10) Message boards : Number crunching : Not getting any tasks, though many are available (Message 42208)
Posted 17 Apr 2020 by doug
Post:
Hi,

I haven't been getting any work from lhc@home for the last nearly week or so, even though the Server Status page has pretty consistently shown 3000+ "Tasks ready to send". I have all apps selected in my preferences. Windows 10, Intel Core i5-3470 with 16GB memory, 4 cores.

When I first joined lhc@home a few weeks ago, I periodically got a few tasks and worked them off (though the BOINC-estimated times for the tasks were way off). But now I'm just not getting any tasks. I've even suspended all my other projects before doing an Update (which usually works with other projects), but that didn't get my any tasks either.

Is it possible that, though there are plenty of new tasks, that none of them are for Windows/Intel machines? If not, why am I not getting new work?

Thanks.

Doug



©2020 CERN