| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home
Thread: Preempted at 100% |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 31
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello, probably my first post here...
I've just noticed one my FAAH (Boinc) WU being preempted at 100% just after having checkpointed. Is it a known problem? (I have not found it in posts from recent months.) Previously I knew it from Rosetta, the devs claim the actual version 5.41 having this issue corrected. There is actually no problem with loosing some work (it is checkpointioing each 30-40 minutes here), but having pretty much of RAM/swap locked if running with "leave in memoty" option on. Especially on memory-hungry projects like FAAH and Rosetta  And even more when my WCG accumulates much negative STD when running long WU with low ressource share. My Rosettas took often a day to start again and immediately go away.(WinXP, Boinc 5.7.5, FAAH 5.28, faah1044_d254n300_x1AJX_03_1) |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
This is a bug in the BOiNC 5.7.x alpha's 'improved' checkpoint / pre-empted project switching needing reporting at BOiNC dev forum. If u suspend the project to which the scheduler switched, it will usually finish the 100% check-pointed work unit.
----------------------------------------
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Dec 8, 2006 6:26:54 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I could believe you (and will try to report it there), but it behaves exactly as Rosettas did with various older Boinc version, as I was observing it maybe last half year. I suppose suspending the following project would not help. In the mean time various other apps got their turn (Seti, Seti Beta, Rosetta, Spinhenge, Einstein) and WCG's deep STD slooooowlyyyyy climbs towards zero...
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Whether the suspending of the subsequent project will make it return to the pre-empted project is very much dependent on how many projects u have and the weight, overworked, debt etc. In the sighted example, it did return for me and after determining the behavior just let it run it's course.... it would always complete it within second after return.
----------------------------------------It's always possibly to have a regressive bug. Only observed this in the alpha's and with HDC particularly. WCG recommends to only use the latest official release, not alpha's.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
If you are not using the alpha, then it is possible that the 100% completion is not accurate. Because of the way the estimation works, sometimes the % complete will appear to "stick" at a particular point while the crunching catches up. Sometimes it even jumps backwards a little.
If you are using the alpha, then Sekerob is undoubtedly correct. We can't officially support the alpha versions here, but a few of us keep up to date with the development process (and contribute to development and alpha testing ourselves). Sekerob, I believe you mean "cited". ;-) |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
U english need to fix your language..... Italian does not have this phonetic ambivalence..... but cited from citation was not meant, rather sighted from seen or observed :O
----------------------------------------On the other hand, you could get cited for using the wrong BOiNC version. Funny was that i wrote decease in one post and it took 3.5 months and a Dutchman to figure out that, that 'fighting decease' could also be 'fighting disease' :D
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, grammatically (syntactically and semantically), "cited" would have been correct. If it wasn't the word you wanted then "In the example shown" or "In the given example" would have been better.
It's a weird language and no mistake. How's that for idiom? :-D |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7849 Status: Offline Project Badges: 
|
It is always interesting to read Sekerob's posts. His command of english is very good as well as his well known penchant for the pun. One does not want to make him feel self conscious by correcting his perceived misuses of english. Generally his meaning is clear although sometimes it takes some effort due to some convoluted sentence structure and homophone usage. You make me think and give me a smile.
----------------------------------------Cheers Sgt.Joe Minnesota Crunchers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm sure he'll tell me if I get annoying. Pedantry is all very well, but one has to draw a line somewhere.
To speak a fraction of the languages he does is a major achievement. Over on UF, 99% of people prefer to have their typos corrected. The other 1% sometimes get a little... irked. So we have to respect their desire to treat language how they want. Anyone trying a bit of l33t speak, however... they are unmercifully mocked. :-) |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Sekerob, I believe you mean "cited". I've noticed that the thread's subject has slowly drifted away...  BTW I'd call the example "mentioned". But my opinion sure do not apply, as I'm everything but naturally english speaking.Never mind. Back to the topic: ...it is possible that the 100% completion is not accurate. Because of the way the estimation works, sometimes the % complete will appear to "stick" at a particular point while the crunching catches up. Sometimes it even jumps backwards a little. In this case, the estimation was very fine. It finally took some 13 seconds to load the data into memory and finish the WU: 08.12.2006 13:48:35 Restarting faah 528 08.12.2006 13:48:40 checkpointed 08.12.2006 14:29:24 checkpointed 08.12.2006 15:09:11 checkpointed 08.12.2006 15:53:27 checkpointed 08.12.2006 15:53:27 Preempting faah1044_d254n300_x1AJX_03_1 09.12.2006 21:51:00 14:21:20 100% STD -9511.24 LTD 09.12.2006 23:19:00 14:21:20 100% STD -12157.57 LTD -55336.81 10.12.2006 02:48:00 14:21:20 100% STD -11640.82 LTD -54883.02 10.12.2006 13:14:00 14:21:20 100% STD -10086.95 LTD -53516.03 10.12.2006 16:49:00 14:21:20 100% STD -9550.83 LTD -53042.39 10.12.2006 17:49:53 Restarting faah1044_d254n300_x1AJX_03_1 using faah 10.12.2006 17:49:59 checkpointed 10.12.2006 17:50:17 Computation finished 10.12.2006 18:56:00 14:21:33 100% STD -0.0 LTD -52766.00 But the additional 2 days to wait... If you are using the alpha, then Sekerob is undoubtedly correct. We can't officially support the alpha versions here, but a few of us keep up to date with the development process (and contribute to development and alpha testing ourselves). Didactylos, I for sure understand this. I've posted about in the boinc_alpha list, maybe I should have better posted it into boinc_dev (I know you are subscribed there), possibly some dialog would came out of it. I personally do not believe it is because of some differences in current alpha. I've finished only a small number (maybe 8) of WUs, so the possibility of noticing such case previously was rather small. But currently the "black sheep" (or "black Peter", as we use to say ) is traveling between the project (it's clearly the alpha client) and the client (it's sure the project) people, and the winner is...  (......) the user  |
||
|
|
|