| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: Always Goes Back to About 68% [RESOLVED] |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 17
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OK,
<leave_apps_in_memory>0</leave_apps_in_memory> The zero indicates that LAIM is not on. You can open the local Computing Preference screen, Disk & Memory tab to verify that the box is not ticked. As you can see in my log, with the extra <cpu_sched> log flag, if it is active and you suspend a task, the Preempting line says at end: [cpu_sched] Preempting GFAM_x1rr6_hPNP_0019728_0143_1 (left in memory) Change that LAIM setting, and you will no see any more retreats and loss of progress when suspending tasks (long as you don't power cycle, boot, or unload the BOINC core client daemon. --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I went into my preferences, and it wasn't on!
---------------------------------------- I put it back on, and suspended a process. The result:Thu May 10 10:26:19 2012 | World Community Grid | task E207501_566_C.28.C19H11N7SSi.01691539.3.set1d06_0 suspended by user Thu May 10 10:26:21 2012 | World Community Grid | task E207501_566_C.28.C19H11N7SSi.01691539.3.set1d06_0 resumed by user Thu May 10 10:26:22 2012 | World Community Grid | Resuming task E207501_566_C.28.C19H11N7SSi.01691539.3.set1d06_0 using cep2 version 640 in slot 7 There is no "left in memory," but it says "Resuming" instead of "Restarting," so I think it's fixed. Thank you so much! :) [Edit 1 times, last edit by Former Member at May 10, 2012 5:27:30 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Wundervoll! The Left in Memory line is as noted in my post only printed when the <cpu_sched> has been added to the configuration file.
Finally, if you open the opening post and edit in [RESOLVED] as I did to this post's title, we can let this one sink till the next one finds it with the same issue. Happy "resume" crunching. --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thank you so much!
 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Glad that this could be sorted out.
Best wishes from Your Harvard CEP team |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just a tip: I had this issue often when running several CEP2 tasks simultaneously, even with LAIM on. This never happened again after I installed more physical RAM. After that I could let run all 8 cores without manual intervention and losing progress. Linux machines seemed to work CEP2s pretty inefficient for me. Everything was fine with my 2 GB 2-threads and my 8 GB 8-threads Windows systems.
But for the beginning, it's surely best to start with the default option 1 task per machine. Best wishes. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, it is correct that if you run a number of CEP wus simultaneously on a machine with little RAM, you may not have enough physical memory to actually make use of LAIM. Each instance should not need more than 500MB of RAM, i.e., if you run on 8 cores you should have 4GB.
Best wishes Your Harvard CEP team |
||
|
|
|