| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: Rather Long MCM1 WUs lately |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 39
|
|
| Author |
|
|
rbotterb
Senior Cruncher United States Joined: Jul 21, 2005 Post Count: 401 Status: Offline Project Badges: 
|
I've got a couple MCM1 WUs that are crunching along - one at 16.5 Hours at 51% done, another at nearly 13 hours at 37% done. That's pretty big even for my old laptop. I run CEP2 WUs from time to time and they run pretty consistently at 12 hours for each WU, but these MCM1 WUs are going to be running all weekend at the pace they are going.
I've seen some partial postings on settings - does this new project need a few settings set a certain way to run well (and shorter)? Otherwise I can keep crunching on these puppies and see what day they finally get done..  |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello rbotterb,
CEP2 has an arbitrary cutoff at 12 hours. Otherwise, it would run much longer. Originally WCG tried to size units at 7 or 8 hours, but that has crept down over the years and FAHV is so short it gives the server conniptions. The staff said that initially they were having difficulty estimating the size of the MCM1 work units that they were creating. We shall just have to see if MCM1 becomes less variable in duration over time and, if so, what the average time settles at. Lawrence |
||
|
|
Warped@RSA
Senior Cruncher South Africa Joined: Jan 15, 2006 Post Count: 425 Status: Offline Project Badges: 
|
I have noticed a big jump in run times for work units received in the last couple of days.
----------------------------------------They seem to be taking about three times as long as those received previously.
Dave
 |
||
|
|
OldChap
Veteran Cruncher UK Joined: Jun 5, 2009 Post Count: 978 Status: Offline Project Badges: 
|
Looking at the average runtimes over the last few days I would guess that the current batches are still being tested for just that: runtime.
----------------------------------------Given the disparate hardware in use to crunch these, the guys building the batches of work probably still need to find the best size to build in order to satisfy the criteria of not hitting the servers too often and still allowing slower hardware to complete in a reasonable time.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My read of what the techs have written: [at this time] No control at all over duration. 5 seconds jobs I've had and 21 hours [v7.26]. Under v7.24 had a 51 hour job... wherever the search space leads. Had tasks that have done 100+ checkpoints in 4-5 hours, had tasks doing 1 checkpoint in many hours, all going valid. Maybe very large production stats will 'reveal' some pattern, but given the checkpoint count variability in a task...
Yesterday the project showed 4.6 hours mean, but given the lapse between issue and validation, will be several more days to see if this was a batch spike or even higher averages are in store. |
||
|
|
grumpygrampy
Senior Cruncher USA Joined: Oct 8, 2011 Post Count: 223 Status: Offline Project Badges:
|
I had 8 wu's yesterday that ran 12 - 16 hrs each. Half of them have now validated, the others pending.
These were all in the same batch MCM1_0000214-xxxx . Glad now i let them run.  |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
Yeap, MCM WUs are getting longer... the evidence (more runtime, less results returned):
----------------------------------------Statistics Date........ Total Run Time......... Points Generated..... Results Returned 11/23/2013....... 152 :187:22:19:09....... 235,282,846.............. 290,206 11/22/2013....... 148 :000:02:49:37...... 210,450,944.............. 345,317 11/21/2013....... 140 :181:00:54:30...... 195,278,837.............. 360,860 CJSL Gotta keep crunching, there's a world to save !!! |
||
|
|
Mumak
Senior Cruncher Joined: Dec 7, 2012 Post Count: 477 Status: Offline Project Badges: 
|
I had one running for nearly 24 hours.
----------------------------------------MCM1_ 0000218_ 1754_ 0-- Time: 22.71 / 23.37 Points: 90.4 / 90.4 Wingmen runtime: 14.66 Only 90.4 points for such a long task ? Another task there took 3.73 hours and got 93.9 points.  [Edit 1 times, last edit by Mumak at Nov 24, 2013 3:33:32 PM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7745 Status: Offline Project Badges:
|
My read of what the techs have written: [at this time] No control at all over duration. Mine are varying all over the place similar to what SekeRob has experienced. I suspect some will go well over 20 hours on one of my Xeons, but they are all still incrementing time and showing progress. No stuck units and the uploads and downloads are going smoothly. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm getting anything from 2+ hours runtime to over 21 hours just now.
I dont mind either way though, as long as they are ok. |
||
|
|
|