Message boards :
Number crunching :
Long WU's
Message board moderation
Author | Message |
---|---|
Send message Joined: 9 Jan 08 Posts: 66 Credit: 727,923 RAC: 0 |
Just got a few work units with estimated 60 hours of completion. Pretty heavy ones. Don't know if the deadline should maybe be more than a week, since you can grab 4 of them per core. So that is 10 days of computation. But lets see how long they actually take. Normaly my WU's on this machine takes 4-5 hours only. |
Send message Joined: 1 Sep 04 Posts: 57 Credit: 2,835,005 RAC: 0 |
Got one that will take 40 hrs to complete. |
Send message Joined: 4 May 07 Posts: 250 Credit: 826,541 RAC: 0 |
I just got one too. wlxu7_nuebb1_8_s_64.31_59.32_10.5_11_7_24_1_sixvf_boinc5503_1 I'll try to keep an eye on it to see what the Remaining (estimated) looks like once it starts running Tom |
Send message Joined: 28 Jan 10 Posts: 8 Credit: 8,635,584 RAC: 0 |
I've got two runing. 9,5% after 4H 37mn 7,7% after 3H 52mn with a Xeon E3 1245 3.3GHz |
Send message Joined: 24 Oct 04 Posts: 1172 Credit: 54,702,229 RAC: 12,625 |
Yes I got quite a few of those too. Just checked all of my hosts and saw the different versions (pni and sse3) Of course the Boinc Manager is doing a "estimation" for the run times. But no doubt these are the longest LHC tasks ever. I have a few running and they are estimated between 40hrs to 52hrs But one of the ones estimated to be over 40hrs is already dropped down to being estimated to be just over 30hrs so we will find these times updated as we run these. The only problem I had was these blocks of long tasks made them run "high priority" ona couple of my hosts which made the T4T tasks stop running so I had to suspend some of the LHC's to get the T4T's to stay running. Didn't have a problem with the Einstein Cuda's though. This sure is a change from the several years of 5 second tasks we got here. I did set up my 2 triple-core hosts to just run LHC's and suspended the Einsteins on those so it can just run these long LHC's Volunteer Mad Scientist For Life |
Send message Joined: 28 Jan 10 Posts: 8 Credit: 8,635,584 RAC: 0 |
One of those big tasks wlxu7 has finiched in 300sec http://lhcathomeclassic.cern.ch/sixtrack/workunit.php?wuid=2468780 |
Send message Joined: 19 Feb 08 Posts: 708 Credit: 4,336,250 RAC: 0 |
I am running one of these long units on my laptop with the AMD APU.It runs high priority so T4T won't start.But I am curious to see what the APU can do, Tullio |
Send message Joined: 1 Mar 07 Posts: 47 Credit: 32,356 RAC: 0 |
I just got one with estimated run time of over 311 hours, but deadline is still 7 days ! http://lhcathomeclassic.cern.ch/sixtrack/workunit.php?wuid=2486449 |
Send message Joined: 9 Oct 10 Posts: 77 Credit: 3,671,357 RAC: 0 |
I got one on my old P4-m 1.4 GHz which has been estimated for 307 hours, but I'm pretty sure it won't take that long to complete. I've seen a lot of wrong estimations in the past on this project ... |
Send message Joined: 3 Oct 06 Posts: 101 Credit: 8,994,586 RAC: 0 |
I just got one with estimated run time of over 311 hours, but deadline is still 7 days! + 1! The size of tasks is increased 10 times! by the same deadline! What is happening? Is the project team trying to repeat the mistakes of pre-alpha, alpha projects? Or... Is the project team trying to set the project to second RICH PEOPLE CLUB after GPUGRID? |
Send message Joined: 17 Feb 07 Posts: 86 Credit: 968,855 RAC: 0 |
I have 3 of them, one finished in 59.12 minutes, but won't upload but that is another problem. One has run 01 hour and 22 minutes and need another 77.39 hour to complete. The last on has done 15 minutes and 80 hours and 17 minutes to go. But uploading will be a problem after they finished. It seems to be a F-Secure thing but no one has a solution for it. Greetings from, TJ |
Send message Joined: 4 May 07 Posts: 250 Credit: 826,541 RAC: 0 |
I just got one too. Something funny is going on. The above Task completed and was verified but with only 170.+ CPU seconds and 0.91 credit. BTW, it looks like all the new Tasks I've been getting are very long. |
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Sorry I should have issued a News. I shall try and send a lot of info tomorrow. As everything is running so well from my side I am trying some 10 million turn jobs....ten to hundred times the usual. BUT at some amplitudes the particles will be lost after just a few turns. I don't know if any will survive for 10 million turns and also floating-point error accumulation may invalidate the results. This is an experiment, a first time ever, and may also show up other issues. Eric. |
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
.and sorry the running time should be a maximum of ten times a one million turn case i.e I estimate 20 to 48 hours max but much much longer on an old Pentium 3 Eric. |
Send message Joined: 22 Oct 07 Posts: 27 Credit: 808,821 RAC: 0 |
I think you should have extended the deadline time. ++PLS |
Send message Joined: 3 Oct 06 Posts: 101 Credit: 8,994,586 RAC: 0 |
trying some 10 million turn jobs....ten to hundred times the usual PLEASE! Correct the deadlines too, because the older machines (Pentium 4, Core etc.) will not finish in time such long tasks. |
Send message Joined: 9 Jan 08 Posts: 66 Credit: 727,923 RAC: 0 |
So far i have had a few that ended after a few hours like normal, but at the moment I have one that has been running for 10 hours and 20 minutes and still says 52 hours to completion, so we will see if it survives the long run. Thanks for the update Eric. |
Send message Joined: 18 Sep 04 Posts: 40 Credit: 60,176 RAC: 0 |
I have aborted two of these on my quad-core and still have two running (14 hrs. with still 36 hrs left). I'm happy to allocate half of my resources to the tasks but they cause everything to go into a high-priority frenzy. Warped |
Send message Joined: 19 Feb 08 Posts: 708 Credit: 4,336,250 RAC: 0 |
I am at 84% after 12 hours on my APU. Looks reasonable. Tullio |
Send message Joined: 30 May 08 Posts: 93 Credit: 5,160,246 RAC: 0 |
I don't mind the very long run times so much if the deadline is appropriate and the credit is reasonable. However, I was a little disappointed with task 5426766: Less than 8 credits/hour after 28 hours on a 1090t... :-( |
©2024 CERN