| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Cure Muscular Dystrophy - Phase 2 Forum
Thread: Parents, children, grandchildren WUs - how does it work? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 76
|
|
| Author |
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
This combo was novel to me... 1 not getting to 60% complete and the second not managing to do the remain 40% in the additional 6 hours permitted:
----------------------------------------CMD2_ 0325-HSP7CA.clustersOccur-1YWM_ A.clustersOccur_ 14_ 108644_ 110297_ 1-- 614 Valid 5-2-10 23:17:38 7-2-10 01:35:59 12.01 251.8 / 229.2 CMD2_ 0325-HSP7CA.clustersOccur-1YWM_ A.clustersOccur_ 14_ 108644_ 110297_ 0-- 614 Valid 5-2-10 23:16:05 9-2-10 17:34:22 6.06 110.1 / 122.1 < moi
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges: 
|
@Sekerob: Your wingman appears to have had one of the very rare Type E situations as per my diagram earlier in this thread [Nov 23, 2009 5:33:20 PM], while you got a Type F.
----------------------------------------Nowadays, I only crunch HCMD2 on 1 of 2 cores of my old 2.5GHz AMD, because it's performance is near the fleet average, so that there are many devices in its performance class and device-matching works reasonably well. Device-matching is abysmal with my Intel 3.25GHz and 4GHz non-HT quads, and too much of their crunching effort would be wasted on HCMD2. [Edit 1 times, last edit by Rickjb at Feb 11, 2010 5:35:37 AM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
E, F, it's a rare combo, the one that's able to finish the whole pack in < 6 hours and the wingman not in 12 probably rarest.
----------------------------------------Slow down... it only needs a single position to take hours to blow through the 6 or 12 hour wall. If only the scientists knew what ticks them off, then they could package these separately which might increase the number of jobs that finish the whole set in 1 run. Given that predominately my grants match the claim quite closely, under 10% deviation, and a result of pure averaging of quorum, the waste is as I sense it is not as large as some postulate. What would happen if the 6 were made 7 hours and 12 became 14 WITHOUT changing the structure inclusions? Momentarily I'm running a long series of [grand]children in ranges from 0.2 to 4 hours. Those for a substantial part could get absorbed into a 7/14 hour limit and still leave the project average well below the 5-5.5 hours... it's 4.6 presently and was 4.2 on Sunday. Memory weak, think uplinger said something to that extend a few months ago... to me rather <i>sooner than later</i>.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
PS, Rickjb, I've put a link in FAQ Index to your Nov.23 diagram post. It's no longer editable, but if you feel the need to repost in polished/update, I'll update the link, parked for now under item 6H.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Mysteron347
Senior Cruncher Australia Joined: Apr 28, 2007 Post Count: 179 Status: Offline Project Badges: 
|
I don't intend to be critical, Sekerob - but I couldn't find that link under 6H - and "Human Proteome Folding (HPF2), Credit, Quorum and Result Validation" seems a very odd place to put it...
Meanwhile and [OT] Item 1P "How to Minimize Tasks Progress Loss during Shut Down/Close/Restart/Resume" - your Jan 22, 2007 7:58:18 PM post with the checkpoint highlighting: Er, the dates/times of ALL of the visible checkpoints are the same. Is it the last non-trivial-sized wcg_checkpoint* filecreated - remembering that the default sort-order is the strange "ascending order of the numeric part of the name" scheme imposed by Windows - not necessarily date-order which some people may assume... |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
No worries Mysteron347, seems I did not hit the final save, probably did the preview, looked at something else and then navigated away... I'll make it a semi-permanent parking spot on my next visit to the index :D
----------------------------------------1P... interesting observation. Don't know why the timestamps are all identical, suspecting the file is rewritten entirely from memory each time there's something to log. [guess]. It's an ancient part very few will employ, thus considering to remove that paragraph, also because the last checkpoint is now visible in the 6.10 client and time past since when looking in the BOINC GUI task properties screen. So thanks for checking, at least someone actually reading and advising on corrections needed.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Feb 11, 2010 6:01:13 PM] |
||
|
|
|