| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 24
|
|
| Author |
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18667 Status: Offline Project Badges:
|
I've noticed lately that the typical runtimes on HFCC WU's on my 2.5 GHz quad have basically doubled to about 12-14 hours each (according to Results Status page). Not really a problem but just wondering if this is a new norm, a factor specific to some recent batches or something like that.
---------------------------------------- |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Seeing wide ranges on my exclusive HFCC duo from 3.5 to as much as 16.4 hours, keithhenry. All validate, there maybe being slightly more verification via the inconclusive path it seems. Maybe because quite a few results are considerably longer or shorter than the project mean. The chart for HFCC you find at the next link tracking progress.
----------------------------------------http://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,24775 edit: changed link to www address and adding a short history of times Lapsed-01 World Community Grid 6.10 hfcc HFCC_s2_00704892_s2_0001 11:44:10 (10:41:09) 18-03-2010 07:19 18-03-2010 07:19 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00728771_s2_0000 03:52:07 (03:35:09) 18-03-2010 03:21 18-03-2010 03:21 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00706986_s2_0000 04:47:23 (04:11:18) 17-03-2010 23:29 17-03-2010 23:29 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00698383_s2_0000 11:46:57 (11:04:44) 17-03-2010 19:35 17-03-2010 19:35 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00700968_s2_0001 10:20:18 (09:48:38) 17-03-2010 18:41 17-03-2010 18:41 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00696981_s2_0001 12:10:31 (11:11:46) 17-03-2010 08:21 17-03-2010 08:21 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00677364_s2_0000 13:27:12 (12:22:20) 17-03-2010 07:47 17-03-2010 07:48 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00677729_s2_0001 06:59:20 (06:36:44) 16-03-2010 20:09 16-03-2010 20:09 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00664977_s2_0000 10:56:48 (10:43:58) 16-03-2010 18:19 16-03-2010 18:20 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00635881_s2_0001 09:00:23 (09:48:32) 16-03-2010 13:10 16-03-2010 13:10 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00634555_s2_0001 10:31:17 (10:26:23) 16-03-2010 07:01 16-03-2010 09:42 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00615654_s2_0001 14:24:30 (12:57:40) 16-03-2010 02:49 16-03-2010 09:41 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00612644_s2_0000 11:03:29 (10:03:58) 15-03-2010 19:30 15-03-2010 19:31 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00612169_s2_0001 10:38:33 (09:41:15) 15-03-2010 12:24 15-03-2010 12:24 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00612160_s2_0001 10:48:15 (09:43:43) 15-03-2010 08:27 15-03-2010 08:27 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00605527_s2_0001 14:39:33 (13:26:31) 15-03-2010 01:45 15-03-2010 01:46 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00608884_s2_0000 10:17:44 (09:35:30) 14-03-2010 21:38 14-03-2010 21:39 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00599025_s2_0001 16:43:42 (15:14:46) 14-03-2010 11:21 14-03-2010 11:21 Reported: Ok World Community Grid 6.10 hfcc HFCC_s2_00599955_s2_0000 16:15:06 (14:53:50) 14-03-2010 11:06 14-03-2010 11:06 Reported: Ok
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Mar 18, 2010 6:22:53 AM] |
||
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18667 Status: Offline Project Badges:
|
Thanks Sek. It's amazing how much info you pack into your charts. Makes an old farts head hurt!
---------------------------------------- I reviewed my Results Status again for all valid HFCC's and there's no pattern, such as wu batch, that I can find. I saw run times from almost 5 hours to 15 hours. Yes, they all validate fine though I'm not sure quorum of one buys us that much on this project as inconclusives seem to run about half the normal PV rate on other projects. Still, they validate in time so not something to stress over! I've obviously gotten used to the way the admins like to keep a project consistent and finally got curious enough to wonder after seeing these long run times here. At least now if someone else wonders, they can read this. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
There's no cutting these to size keithhenry. The last 4 on the duo did
----------------------------------------5:31 hours (v 6.10) 12:35 hours (v 6.10) 7:40 hours (v 6.11) 16:50 hours (v. 6.11) The quad meantime did one in 3:30 hours. The DCF is not happy. At low cache it causes work fetches to be quite late... so it's up to 0.9 now. Even on doubling that still makes it to return work under 2 days.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It's strange, on my fast new quad the HFCC WU times are very consistent all at 4:28 at the moment, however on my slower computers there is tremedous variation. E.g., on my laptop one is just finishing at 14 hr, whereas the next one to start estimates 52 hr!
Are my fast and slow machines receiving different types of WU? |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Farpung, are you quoting the TTC times of very consistent at 4:28? Because of the difference in run times between jobs, your client's Duration Correction Factor (DCF) is responding strongly. If there were a series of shorter jobs, of any WCG science, then a sudden long of 14 hours, the client thinks that all following jobs will be long, so it applies this DCF multiplicator. The 52 hours will come down when the task actually runs.
----------------------------------------Not as far as I know is there a general fast devices/slow devices matching [there is for HCMD2 for a reason]. If that were, then I'd send the big jobs to the fast machines and the shorter to the slow. There's development plans to have that actually happen. HFCC is one of those very non-deterministic algorithms. It's no exact predicting as every step is series of docking attempts trying to find the lowest binding energy... the lower the better. If it cant find a certain minimum, it will try until it does, or does not, which is when the next docking step will be tried et cetera.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ah, now I understand!
I thought I saw consistent times on the fast machine because I was looking at the list of waiting WUs that were all assigned the same estimated times. By contrast, the WUs on the slow machines come in one at a time, and in fact, the job I said was estimated at 52 hours will soon be complete after only 27.5 hours. |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
Just switched to this project and mine are 18.5 hrs run time on my mac pro. Wow!
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Bearcat, re your worries about these jobs suffering from similar to HCC, no they don't. The sign is large amounts of kernel time and mountains of PF soft errors AND lots of delta. There isn't. They are just tough jobs, very non-deterministic in nature. Got just now a 4:10 done and currently chewing on a near 20 CPU hour, same duo.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
edit, seems my duo is very good at them... zero redundancy and grants only known in me dreams... the previous previous result, telling the flops are pretty real:
----------------------------------------Project Name: Help Fight Childhood Cancer Created: 19-3-10 Name: HFCC_s2_00831608_s2_0000 Minimum Quorum: 1 Replication: 1 Result Name App Version Number Status Sent Time Time Due / Return Time CPU Time (hours) Claimed/ Granted BOINC Credit HFCC_ s2_ 00831608_ s2_ 0000_ 0-- 611 Valid 21-3-10 04:27:06 22-3-10 09:25:02 13.27 131.8 / 147.6
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|