| 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: CEP2 validator disabled - Resolved/Resumed |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 89
|
|
| Author |
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
We are seeing a slightly higher error rate than normal since we restarted sending out work again, but it is much lower than it was before we stopped it previously. The current error rate is not high enough to stop the project but we are looking into the errors to see what we can do to continue to decrease them. At this point anything we do will not require us to stop the project again, however if that changes we will notify members.
Thanks, armstrdj |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
Thanks armstrdj & cleanenergy for the explanations. So if I understand correctly, "pushing at the boundaries" means "to go boldly where no molecule has gone before"... I'll try to keep that in mind
---------------------------------------- .CJSL Crunching boldly for a better future... PS: apologies to the Trekkies |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
We are seeing a slightly higher error rate than normal since we restarted sending out work again, but it is much lower than it was before we stopped it previously. The current error rate is not high enough to stop the project but we are looking into the errors to see what we can do to continue to decrease them. At this point anything we do will not require us to stop the project again, however if that changes we will notify members. Thanks, armstrdj We are working on some very large basis sets within limited run times. The programs work the time limits are not. Why can't they be very extended or even dropped?? It's just computer code..... what's the problem?? [Edit 1 times, last edit by Former Member at Sep 10, 2014 12:20:12 AM] |
||
|
|
Seoulpowergrid
Veteran Cruncher Joined: Apr 12, 2013 Post Count: 823 Status: Offline Project Badges: 
|
FYI - My settings are to run the CPU 100% of the time whenever my computer is on. The Results Status page says CPU Time 18.00 / Elapsed Time 34.59 but I know I passed at least 24 CPU hours on the WU before I went out for a nice evening. AKA the 18 hour time limit isn't holding (and I am happy as I want to finish the file as much as I can).
----------------------------------------WU E225204_911_S.226.C29H22N2O2.SUCFRAWCOSKMBS-UHFFFAOYSA-N.8_s1_14 Cheers!  |
||
|
|
Yarensc
Advanced Cruncher USA Joined: Sep 24, 2011 Post Count: 136 Status: Offline Project Badges:
|
FYI - My settings are to run the CPU 100% of the time whenever my computer is on. The Results Status page says CPU Time 18.00 / Elapsed Time 34.59 but I know I passed at least 24 CPU hours on the WU before I went out for a nice evening. AKA the 18 hour time limit isn't holding The 18 hour limit is for the CPU time, which is where yours cut off. The fact that your elapsed time is so much higher is a little odd, but it could be from you doing other things on the system at the same time, or partially from running multiple CEP work units at the same time. When you select "100% CPU time" it means that WCG can use *up to* 100%. It is given the lowest priority in windows so that anything else you are running takes priority, this is done so that it doesn't slow down anything you are actively working on. When other programs are using the CPU, the elapsed time keeps going up, but the CPU time does not (since you aren't crunching anything for the 2 seconds the other program is using the CPU) Hope that helps explain the difference and the reasoning. yaren [Edit 1 times, last edit by Yarensc at Sep 10, 2014 2:40:18 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
When you select "100% CPU time" it means that WCG can use *up to* 100%. It is given the lowest priority in windows so that anything else you are running takes priority, this is done so that it doesn't slow down anything you are actively working on That's only partly true. He can set WCG to run 100% and could have his % of CPU set much lower say 50% or less and that would explain a 35 hour computer run time with only 18 hours CPU time. Needs to check his computer profile and custom settings. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
FYI - My settings are to run the CPU 100% of the time whenever my computer is on. The Results Status page says CPU Time 18.00 / Elapsed Time 34.59 but I know I passed at least 24 CPU hours on the WU before I went out for a nice evening. AKA the 18 hour time limit isn't holding (and I am happy as I want to finish the file as much as I can). WU E225204_911_S.226.C29H22N2O2.SUCFRAWCOSKMBS-UHFFFAOYSA-N.8_s1_14 Cheers! The far apart cpu time / elapsed, indicating a 52 percent efficiency is indicative of some other higher priority process that ate the cpu time. Elapsed is the time the agent was allowed to run, per your settings. Of that, only 52 percent could actually be used by the lowest priority science app, which is what they are set to run at, hardcoded. The remaining time went to something else, not boinc. The place to look is task manager, key-combo ctrl+shft+esc to open. Hit "show processes from all users" button at left bottom, then sort the processes by cpu and cpu time. If you can't see those columns, you can add them through view > select columns. These will reveal which are the momentary and cumulative big time users, of the programs still running. On an idle quad system you'd see the wcg processes as using 25 percent or something very close. That number equates to 100 percent of 1 core. The pc runs cep2 at about 97 percent efficiency when idle. |
||
|
|
Seoulpowergrid
Veteran Cruncher Joined: Apr 12, 2013 Post Count: 823 Status: Offline Project Badges:
|
Thanks for the replies.
----------------------------------------I have an i7 870 @ 2.93 GHz, no overclock. It has 8 threads. Boinc settings are "use 100% processors" & "use at most 100% CPU time". I opened task manger and other than a little activity in Chrome, all CPU usage were with Boinc processes. Boinc version 7.2.42, Win 7, 4 gigs of RAM. I don't use this machine for heavy CPU usage outside of Boinc. I could watch the process tick upwards and the elapsed time hit around 24 hours before I left for the evening. As 18 is the cap I took mental note and checked my WCG page the next morning - it said "CPU Time 18.00 / Elapsed Time 34.59". The "elapsed" time in Boinc should be the "CPU time" that WCG "my contribution" page says, right? I understand being off by 10~20 minutes, but this was a it outside that range :) Perhaps my minimal Chrome usage was using the same CPU core as the "18 hour" WU and Boinc was counting it anyways? For my wingman it says it took them "12.01 CPU hours" so guess whoever had the other copy has significantly higher GHz than I do. I feel confused, rather curious, and it looks like it is just me. *shrugs* |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
@Seoulpowergrid,
----------------------------------------If many of your 8 threads were being used for CEP2, it might be possible that 4 GB of RAM was insufficient - did you check RAM usage in task manager at the time? Lots of page swaps would drastically affect CEP2 efficiency. The "elapsed" time in Boinc should be the "CPU time" that WCG "my contribution" page says, right? Wrong. Elapsed time in BOINC Manager is Elapsed time in My Contribution (Results Status)  . In BOINC Manager, select a task and click on the Properties button; that will list both CPU time and Elapsed time for that task at that moment.[Edit 1 times, last edit by Former Member at Sep 12, 2014 2:06:51 PM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
Something has changed (check your results). All the "errors" have been eliminated on my results, and for at least the past 24 hours I have been getting mostly a zero-quorum.
They must have figured out how to distinguish the real machine errors from just the algorithmic problems noted above. Maybe they just remove from the list the ones where everyone errors out? |
||
|
|
|