log in
1) Message boards : Number crunching : Native Python v2.7 Application v1.05 (sse2) - waiting/suspended --> still running (Message 2744)
Posted 8 Jan 2014 by Profile 7ri9991 [MM]
OK. Thanks.
2) Message boards : Number crunching : Native Python v2.7 Application v1.05 (sse2) - waiting/suspended --> still running (Message 2741)
Posted 3 Jan 2014 by Profile 7ri9991 [MM]
On one of my Gentoo Linux machines, I was wondering why my PC with 8G of Ram was maxing out and then maxing out my swap partition. After killing about 15 instances of python2.7.exe, I realized that these processes don't stop once they're started. The only thing that stops them is finishing, being aborted or the kill command. If you suspend 4 of them and then 4 more start, then you have 8 of them running, eating up clock cycles and ram. Suspend MindModeling and they're still running while a project that's supposed to be running is also taking clock cycles and ram.

I'm guessing that there's a problem with the wrapper.

Anyways, if you're running more than one project eventually MM tasks will go to "waiting" while another project gets its turn. I only have enough ram for 4 tasks not 8, but the "waiting" tasks are still running.

I guess if I ever want to reach 100k, I need to just run this project by itself.
3) Message boards : Number crunching : Pypy Application Errors on Windows (Message 2736)
Posted 29 Dec 2013 by Profile 7ri9991 [MM]
I just noticed this thread, but I just noticed also one of my PC's locked up due to a lack of memory. When I looked for the reason, I had a process 'pypy' taking up 3G of memory. That process came from mindmodeling and I wasn't even running a mindmodeling task. Is it possible to just let Linux hosts use their own python install? I don't need pypy zombies taking up resources on my desktops.
4) Message boards : Number crunching : Server aborting started tasks?! (Message 2592)
Posted 11 Aug 2013 by Profile 7ri9991 [MM]
Sorry, trigggl. We cancelled our worknits because we were getting a lot of failures from long-running tasks and we wanted to prevent any more wasted resources. But since we also understand that aborting running tasks is detrimental to our volunteers, we slightly modified BOINC's "cancel workunit"'s feature to only cancel unsent work. In the future, this will give the work that's out there a chance to complete, even if it may be unlikely.


Cool. Thanks Tom!

Fortunately, I haven't seen much of a problem with the tasks.

Shouldn't there be a way to cancel the tasks that haven't started yet without affecting the ones already started?
5) Message boards : Number crunching : Server aborting started tasks?! (Message 2588)
Posted 9 Aug 2013 by Profile 7ri9991 [MM]
That's one of the big no-no's of boinc'ism. The server shouldn't be aborting tasks that the client has already started. I just lost 3 hours of work and probably more.
6) Message boards : Number crunching : Linux tasks on AMD (Message 2026)
Posted 6 Jun 2012 by Profile 7ri9991 [MM]
Has anyone else noticed tasks crashing all other tasks on an AMD Linux system? I got some work on one of mind and it seems to have immediately killed my Donate tasks as well as some suspended simap tasks. My intel machines seem unaffected.

Never mind: The system was getting unstable. Don't know why.
7) Message boards : Number crunching : Credit New? (Message 2006)
Posted 10 May 2012 by Profile 7ri9991 [MM]
I notice that the first few tasks on each host was netting around 12 credits, then each one backed off to roughly 3. Is this a credit new-ism?

I notice the server software is updated.
8) Message boards : Number crunching : current jobs - not getting any (Message 1936)
Posted 24 Jan 2012 by Profile 7ri9991 [MM]
Are the tasks all out for Change-Signal-Test-Sweep already?
9) Message boards : Number crunching : More tasks issued (Message 1935)
Posted 22 Jan 2012 by Profile 7ri9991 [MM]
Well, that was a good size helping. Managed to get 5k out of it.
10) Message boards : Number crunching : compute errors on one host only (Message 1932)
Posted 20 Jan 2012 by Profile 7ri9991 [MM]
I'm getting consistent compute errors on one host only.

http://mindmodeling.org/beta/result.php?resultid=155539

The first few tasks were good, but all successive ones are getting compute errors. This one is kind of buggy with the boinc client, so maybe this problem is related and can help me find what's causing this host to be a little off.
11) Message boards : Number crunching : More tasks issued (Message 1927)
Posted 18 Jan 2012 by Profile 7ri9991 [MM]
I'm frustratingly close to 10k.
12) Message boards : News : Faster Download Speeds (Message 1875)
Posted 29 Jun 2011 by Profile 7ri9991 [MM]
It will probably be at least several weeks before a significant number of workunits are generated, but there may be some occasional small jobs in between.

So, what we just had, is that considered a significant number or a small job?
13) Message boards : News : Faster Download Speeds (Message 1863)
Posted 28 Jun 2011 by Profile 7ri9991 [MM]
My downloads are 4k just like in May.

I do appreciate the effort, though.
14) Message boards : News : Site Updates (Message 1804)
Posted 25 Feb 2011 by Profile 7ri9991 [MM]
My birthday is on the 25th. So, in case you didn't know what to get me...

So, how about it?
15) Message boards : News : Site Updates (Message 1803)
Posted 23 Feb 2011 by Profile 7ri9991 [MM]
You can expect some relatively small batches of workunits within the month, but larger jobs are TBD.

My birthday is on the 25th. So, in case you didn't know what to get me...
16) Message boards : Number crunching : One batch of work, please - hold the screensaver (Message 1526)
Posted 29 Dec 2010 by Profile 7ri9991 [MM]
Seriously, we're not delaying an app because of a screensaver are we?
17) Message boards : Number crunching : Is MM weeks months or years away from starting back up full time? (Message 1521)
Posted 3 Dec 2010 by Profile 7ri9991 [MM]
The best way IMO to post the updates is with the news feed. Those will get noticed by a lot more people. News feeds get updated in vb forums that a lot of teams use.
18) Message boards : Number crunching : negative points on Free-DC (Message 1510)
Posted 12 Jul 2010 by Profile 7ri9991 [MM]
Credits are back now. Thanks!

ditto
19) Message boards : Number crunching : negative points on Free-DC (Message 1506)
Posted 7 Jul 2010 by Profile 7ri9991 [MM]
Did we just lose credits?
20) Message boards : Number crunching : stats cron job (Message 1492)
Posted 16 Apr 2010 by Profile 7ri9991 [MM]
I made some permissions modifications to the stats folder. That should fix the problem; I'll double-check BoincStats when they update next.

It's showing up on Free-DC now. Thanks!


Next 20

Main page · Your account · Message boards


Copyright © 2019 MindModeling.org