| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: Task Hanging at 51.760% |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 17
|
|
| Author |
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges: 
|
Another thought comes to mind. What OS are you using, how many tasks do you have running at the same time, and how much memory do you have ? I have this problem as well, recently more than in the previous months, but this is (at least for me) a problem that goes back years.If you are short on memory for the number of tasks you have, that will cause the computer to churn as it will be using constantly using its swap space. If you are using Windows, have a look at task manager to see your memory use. If on Linux use System monitor. If you are maxing out your memory, try running fewer tasks to see where your sweet spot might be. Cheers Being hung in the 50%s is for me rather the exception, usually it "hangs" in the high 90s or even at 100%. And this while other jobs finish just fine. I thought that this could be memory related, but this happens across a lot of different hosts, at different times, with varying degrees. But the memory use of MCM1 WUs is supposed to be ony 400MB, and this happens on machines that have 16GB and more, haven't checked the actual memory usage at the time I abort them. And restarting the host is in most cases not an option as I don't run "a farm" but all machines are working hosts. And while most on my hosts are Windows, from XP to Windows 10, this happened to me on macOS and Linux as well, so it isn't a directly OS related issue... Ralf |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
But I do notice this new Boinc manager compared to the old one on my now defunct Win 7 PC. Back then each work units is its own process, but now all workunits is a child process under BOINC manager. That's only a default view issue of the task manager in Windows 11, it has nothing to do with how the memory management under Windows 11 works, it isn't any different from Windows 10 (or 8.1,...)Ralf |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges:
|
Another thought comes to mind. What OS are you using, how many tasks do you have running at the same time, and how much memory do you have ? I have this problem as well, recently more than in the previous months, but this is (at least for me) a problem that goes back years.If you are short on memory for the number of tasks you have, that will cause the computer to churn as it will be using constantly using its swap space. If you are using Windows, have a look at task manager to see your memory use. If on Linux use System monitor. If you are maxing out your memory, try running fewer tasks to see where your sweet spot might be. Cheers Being hung in the 50%s is for me rather the exception, usually it "hangs" in the high 90s or even at 100%. And this while other jobs finish just fine. I thought that this could be memory related, but this happens across a lot of different hosts, at different times, with varying degrees. But the memory use of MCM1 WUs is supposed to be ony 400MB, and this happens on machines that have 16GB and more, haven't checked the actual memory usage at the time I abort them. And restarting the host is in most cases not an option as I don't run "a farm" but all machines are working hosts. And while most on my hosts are Windows, from XP to Windows 10, this happened to me on macOS and Linux as well, so it isn't a directly OS related issue... Ralf Since this has never happened to me, I can only speculate on a possible cause. One thought which comes to mind, since these are working machines, is a possible conflict with one of the existing applications. Depending on the number of applications running, this could be devilishly hard to pin down, especially since it does not happen to all the jobs running. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Since this has never happened to me, I can only speculate on a possible cause. One thought which comes to mind, since these are working machines, is a possible conflict with one of the existing applications. Depending on the number of applications running, this could be devilishly hard to pin down, especially since it does not happen to all the jobs running. The only catch with that theory is that this also happens when a host is not actively being used, with no other open applications, like over long weekends, holidays, or "my" CAD machine next to my desk, which isn't being used, as it is more of a backup/spare workstation here in the office, for +95% of the time.Cheers And the thing that puzzles me and which makes it so hard to track down is that at the same time that one specific WU is "hung", plenty of other WUs, even from the same batch, are working just fine...  Ralf |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges:
|
Since this has never happened to me, I can only speculate on a possible cause. One thought which comes to mind, since these are working machines, is a possible conflict with one of the existing applications. Depending on the number of applications running, this could be devilishly hard to pin down, especially since it does not happen to all the jobs running. The only catch with that theory is that this also happens when a host is not actively being used, with no other open applications, like over long weekends, holidays, or "my" CAD machine next to my desk, which isn't being used, as it is more of a backup/spare workstation here in the office, for +95% of the time.Cheers And the thing that puzzles me and which makes it so hard to track down is that at the same time that one specific WU is "hung", plenty of other WUs, even from the same batch, are working just fine... Ralf Well, it was only a theory. Possible lock at the register level ????? If so, I would have no idea how to track that down. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Since this has never happened to me, I can only speculate on a possible cause. One thought which comes to mind, since these are working machines, is a possible conflict with one of the existing applications. Depending on the number of applications running, this could be devilishly hard to pin down, especially since it does not happen to all the jobs running. The only catch with that theory is that this also happens when a host is not actively being used, with no other open applications, like over long weekends, holidays, or "my" CAD machine next to my desk, which isn't being used, as it is more of a backup/spare workstation here in the office, for +95% of the time.Cheers And the thing that puzzles me and which makes it so hard to track down is that at the same time that one specific WU is "hung", plenty of other WUs, even from the same batch, are working just fine... Ralf Well, it was only a theory. Possible lock at the register level ????? If so, I would have no idea how to track that down. Cheers And I doubt it is a direct memory issue of the system. I had first thing this morning checked on a machine in the conference room, which is used only for occasional Zoom meetings and presentations, still running Windows 7/64bit. I had rebooted that machine just yesterday, so no other application has been running since and yet, there were 3 hung tasks with +12h runtime, 2x100% and not remaining time and one stuck at 99.31% with 16min to go (but not changing). And this is an i7 with 4 cores/8 threads and 16GB of RAM. If the 400MB/WU memory requirement is still valid, that would mean that WCG would use only less than 4GB out of 16GB of that machine. So counting in OS and things like AV being loaded after a restart, that would have been another 8GB to spare... Ralf |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges:
|
One of my machines is an I7-3770 running Win7-64. It is a general purpose machine running a variety of other applications. No problems.
----------------------------------------It would be interesting to have more different work units to crunch and see if the condition affects more than one of them. If it can be pinned down to only one kind of work unit, that would indicate some kind of software bug and almost entirely eliminate a hardware issue. Perplexing issue. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
|