Message boards :
Number crunching :
4.3 scheduler stops/restarts same wu in same second? each hour?
Message board moderation
Author | Message |
---|---|
jshenry1963 Send message Joined: 17 Nov 04 Posts: 182 Credit: 68,878 RAC: 0 |
Does anyone know why the schedule would: start wu a proj a stop wu b proj b stop wu a proj a start wu b proj b It does this within the same second. Log is shown below Occurs once an hour 5/12/2005 3:33:04 PM|ProteinPredictorAtHome|Restarting result h0007A_1_54854_0 using mfoldB125 version 4.28 5/12/2005 3:33:04 PM|SETI@home|Pausing result 19dc04aa.8271.1072.215900.154_6 (removed from memory) 5/12/2005 3:33:04 PM|ProteinPredictorAtHome|Pausing result h0007A_1_54854_0 (removed from memory) 5/12/2005 3:33:04 PM|SETI@home|Restarting result 19dc04aa.8271.1072.215900.154_6 using setiathome version 4.11 I have it set up to do the schedule checking once per hour. Running S@H, P@H, E@H, CPDN, with S@H at highest %. (TMR test seti, normal boinc 4.38) Thanks, and Keep on crunchin' John Henry KI4JPL Sevierville TN I started with nothing, and I still have some of it left. <img src="http://www.boincstats.com/stats/banner.php?cpid=989478996ebd8eadba8f0809051cdde2"> |
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
Hi John, Is this a single occurance? if so then you might have done a manual project update which caused it to switch. I'm not sure. or maybe a doubleclick? I'll be driving past Sevierville sometime Saturday afternoon. Wave if you see me. lol [edit] just saw the each hour, so then, I've just wasted our time. nevermind. sounds like a bug John, somewhere I saw a thread where a user reported that his client would crunch for 60 min, then stop for 60, then start back up. User was using only one project. I don't know if this might be related, and I can't seem to find the thread. Hang on, maybe it was on Einstein forum. brb check this out |
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
John, look at the 4.39 thread people are reporting double requests there. The description is different, but what might be happening to you might be related if there was a call to "Reschedule" and that call stops one project and starts another (as I have seen it do), but the double hit caused you to stop/start twice. tony |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
Are your resource shares even? I've never sat and watched when switchover happens on a project with a larger resource share. Does it not even try to switch, or does it do something like you've seen? Because basically, it paused then restared S@H. Why it got Predictor in the mix, I couldn't say. Perhaps Predictor is next inline for CPU time? Just a guess here... |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
<blockquote>Hang on, maybe it was on Einstein forum. brb check this out</blockquote> But since John is using a HT CPU, he can't got for the one minute switch between applications. Or I don't think so, not if he has more than two projects going. I already left the problem on the developers list, but no answer there yet. -( |
jshenry1963 Send message Joined: 17 Nov 04 Posts: 182 Credit: 68,878 RAC: 0 |
Still continue to get this type of a rescheduling, every hour, once every few hours it actually does do a reschedule to a different work unit, but continuing to get the proj a unit a paused proj b unit b restarted proj b unit b paused proj a unit a restarted. Continues every hour on the hour. I don't see it on any other machine I run. Maybe I'm crying over drops of spilled milk, but enough drops of spilled milk will make a gallon. (sorry for the worldy, a litre) Resource shares are not even, 100, 10, 10, 10, 10. So I do understand it keeping most of everything to resource A, but it shouldn't stop it, start adifferent one, and then stop/restart so that A is going again. Wasteful, and potential times of disk or other errors causing bad computing. Thanks, and Keep on crunchin' John Henry KI4JPL Sevierville TN I started with nothing, and I still have some of it left. <img src="http://www.boincstats.com/stats/banner.php?cpid=989478996ebd8eadba8f0809051cdde2"> |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
If A has a share of 100 and B has a share of 10, then it will run A for 10 hours, switch to B, and run that for an hour, then switch to A and run that for 10 hours ... With A having a share of 100, and B, C, and D having a share of 10 each, then there will still be cycles, but they will be more complicated. If you are worried about wasting processing time etc, change your switch time to 3 or 4 hours from 1. Make certain that the results stay in memory when paused. Both of these are general settings. BOINC WIKI |
jshenry1963 Send message Joined: 17 Nov 04 Posts: 182 Credit: 68,878 RAC: 0 |
Hi John, I'm not really worried about the time it takes for the two switches, I was just mentioing that it is occurring. It could be a bug that is known, if not, it is now. I also understand the % share and the switch rate, that is how it is working. Just that every hour, when it is suppose to go from seti to seti, it switches from seti to something else, then back to seti in that same second. Its not an issue really to me, just mentioning it is happening. I know I get the right results at the end of the day. Thanks for your help, Thanks, and Keep on crunchin' John Henry KI4JPL Sevierville TN I started with nothing, and I still have some of it left. <img src="http://www.boincstats.com/stats/banner.php?cpid=989478996ebd8eadba8f0809051cdde2"> |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.