Message boards : Number crunching : To completion times for new tasks
Message board moderation

To post messages, you must log in.

AuthorMessage
Keith T.
Avatar

Send message
Joined: 1 Mar 07
Posts: 47
Credit: 32,356
RAC: 0
Message 18960 - Posted: 7 Feb 2008, 15:31:49 UTC

Just noticed in BoincView that my old P233 MMX box has a new LHC task highlighted in light yellow instead of the usual green.

Then I noticed that my Athlon's latest LHC tasks have an estimated completion time of 8.5 hours instead of the usual 55 minutes.

It looks like we have some new longer tasks to crunch, I hope the old Pentium MMX makes it OK as it currently estimates completing about 4 hours past the deadline!
ID: 18960 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 27 Oct 07
Posts: 185
Credit: 3,297,428
RAC: 0
Message 18961 - Posted: 7 Feb 2008, 15:40:57 UTC

I hope for your sake that the estimates are reasonably accurate!

I got a very few of these wm71_m7 yesterday - they were estimated at ~6 hours on a Q6600, but ran to completion in under 1.5 hours. Today's WUs have the same estimate, but are starting to look as if they'll take their full allocation and then some.
ID: 18961 · Report as offensive     Reply Quote
PovAddict
Avatar

Send message
Joined: 14 Jul 05
Posts: 275
Credit: 49,291
RAC: 0
Message 18962 - Posted: 7 Feb 2008, 15:45:48 UTC - in response to Message 18960.  

I have work with a 6 hour estimate, and so far (1h 13m runtime, 19.5%), it seems accurate.
ID: 18962 · Report as offensive     Reply Quote
para_doks

Send message
Joined: 2 Sep 04
Posts: 9
Credit: 316,683
RAC: 0
Message 18963 - Posted: 7 Feb 2008, 15:57:10 UTC - in response to Message 18960.  

as shown in screensaver graphics they will calculate 1000000 turns. others were probably 100000 turn WU's.
ID: 18963 · Report as offensive     Reply Quote
KWSN Sir Clark
Avatar

Send message
Joined: 24 Jul 05
Posts: 35
Credit: 19,384
RAC: 0
Message 18964 - Posted: 7 Feb 2008, 16:55:59 UTC

Yep......they must be longer.

I've got a WU sitting at 9.7% after 45 minutes


www.chris-kent.co.uk aka Chief.com
ID: 18964 · Report as offensive     Reply Quote
superempie

Send message
Joined: 28 Jul 05
Posts: 24
Credit: 5,792,993
RAC: 6,371
Message 18965 - Posted: 7 Feb 2008, 17:01:39 UTC

I have the same thing overhere.
ID: 18965 · Report as offensive     Reply Quote
Profile John Hunt

Send message
Joined: 13 Jul 05
Posts: 133
Credit: 162,641
RAC: 0
Message 18966 - Posted: 7 Feb 2008, 17:06:58 UTC


I've had two completed in 35 mins each, seven look like six hour jobs. (Q6600 quad)
The remaining one may take 12 hours. (H/T P4 crunching alongside CPDN)


ID: 18966 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 27 Oct 07
Posts: 185
Credit: 3,297,428
RAC: 0
Message 18969 - Posted: 8 Feb 2008, 0:34:37 UTC - in response to Message 18963.  

as shown in screensaver graphics they will calculate 1000000 turns. others were probably 100000 turn WU's.

A factor of 10 seems right. Q6600s which were returning the old tasks in 37 minutes are now taking 370 minutes (6 hours 10 minutes).
ID: 18969 · Report as offensive     Reply Quote
Rabinovitch
Avatar

Send message
Joined: 11 May 07
Posts: 23
Credit: 3,061,786
RAC: 344
Message 18973 - Posted: 8 Feb 2008, 3:49:56 UTC - in response to Message 18969.  

A factor of 10 seems right. Q6600s which were returning the old tasks in 37 minutes are now taking 370 minutes (6 hours 10 minutes).


Who knows... My Athlon 64 X2 5200+ usually needs about 30 minutes OR LESS for WU, but now it promises me to finish WU in approximately 6 hours. So, like chinese says, "wei?..".

From Siberia with love, Alexey V. Shcherbina aka Rabinovitch.
ID: 18973 · Report as offensive     Reply Quote
Perle
Avatar

Send message
Joined: 25 Oct 04
Posts: 81
Credit: 42,406,603
RAC: 15,828
Message 18976 - Posted: 8 Feb 2008, 5:38:19 UTC
Last modified: 8 Feb 2008, 5:38:41 UTC

Fun observation.
I use an older Compaq sp750 as a SCSI drive tester.
It is a dual P3 (slot 2) Xeon at 1000mhz fsb 133 L2 cache 256mb.
this system want 9 hours for a wu.

Yet my poweredege6850 with 4 Xeon MP 2.66 processors wants 15 hours.

so curious this is.
ID: 18976 · Report as offensive     Reply Quote
glaesum

Send message
Joined: 16 Oct 06
Posts: 15
Credit: 137,531
RAC: 114
Message 18997 - Posted: 9 Feb 2008, 0:04:13 UTC

my main box has gone from 75-80mins to over 12.5hrs so this agrees with the x10 experience reported.

anyway, I have a different question: both my boxes went down last night - probably some power glitch - one completely switched off and the other got stuck at the log in screen losing a night's work. does 'sixtrack' application have proper save points in the middle of a work unit?

the main pc seemed ok, at least it is reasonably fast if the not quite the latest kit, but the ancient AMD 650mhz athlon on win98 looked like it had to start again from scratch. Since it's predicted to take over 2days to complete the wu I can't exactly afford to get behind crunching to turn it round in time!!! lol
(had to suspend other projects until this one clears)
ID: 18997 · Report as offensive     Reply Quote
KAMasud

Send message
Joined: 7 Oct 06
Posts: 108
Credit: 22,919
RAC: 0
Message 19001 - Posted: 9 Feb 2008, 5:24:28 UTC


Check Points are well placed. Never had any problems even with chronic power failures.
Check your machine. Maybe some other problems. Have you blown the dust bunnies out recently or is there a collection in there?
Regards
Masud.
ID: 19001 · Report as offensive     Reply Quote
glaesum

Send message
Joined: 16 Oct 06
Posts: 15
Credit: 137,531
RAC: 114
Message 19005 - Posted: 9 Feb 2008, 19:42:59 UTC - in response to Message 19001.  
Last modified: 9 Feb 2008, 19:44:33 UTC

Check Points are well placed. Never had any problems even with chronic power failures.
Check your machine. Maybe some other problems. Have you blown the dust bunnies out recently or is there a collection in there?
Regards Masud.

the shutdown must have been a too hard landing for the old win98 pc to cope with; anyway, once back up it cracked on with the work unit knocking it out in 36hrs and well beating the two day forecast. I just wish all the credits from this host would show up in my combined stats - it's been 6wks since I attached, any cpid probs should have cleared by now. (just LHC doesn't show - malaria and rosetta are fine)

as for the dust bunnies - I'm red-faced and guilty as charged! there's lots of dust here and I've have been meaning to unplug everything and clean up for months but can't face it... ;)
ID: 19005 · Report as offensive     Reply Quote
metalius
Avatar

Send message
Joined: 3 Oct 06
Posts: 99
Credit: 8,397,239
RAC: 3,602
Message 19008 - Posted: 10 Feb 2008, 11:30:52 UTC

WU's from LHC are now 10 times bigger by daily quotas 10 WU pro CPU. So, in reality we have no restrictions anymore - this is great!
ID: 19008 · Report as offensive     Reply Quote
Keith T.
Avatar

Send message
Joined: 1 Mar 07
Posts: 47
Credit: 32,356
RAC: 0
Message 19018 - Posted: 13 Feb 2008, 16:11:49 UTC

The old P233 MMX completed the task in time http://lhcathome.cern.ch/lhcathome/workunit.php?wuid=2134134

CPU time = 519688.219 seconds = 144.35783861111111111111111111111 hours = 6.0149099421296296296296296296296 days

Interesting to note that it claimed ~ 1 credit/hour

Result was Valid

Received 13 Feb 2008 15:50:08 UTC
Deadline 14 Feb 2008 6:47:52 UTC

I just tried to find the mimimum reccomended CPU spec for LHC on this website, but I can't find it. I know that most BOINC projects do list this.

Keith
ID: 19018 · Report as offensive     Reply Quote
mitrichr
Avatar

Send message
Joined: 20 Dec 07
Posts: 69
Credit: 599,151
RAC: 0
Message 19021 - Posted: 14 Feb 2008, 14:27:20 UTC

Interesting things going on. I also at first saw WUs showing 100,000 turns, and then a shift to 1,000,000 turns. Now, I just got a WU at 100,000 turns again.

I was told, when I asked about the million turns WU's that they were simulations of what would actually go on in the collider. Is it possible that the hundred-thousand turns are tests of this software, regarding stuff like "crashing into walls" errors?

Or, if anyone knows what is going on, I, for one, would like to know. I mean, I am just happy that this project is active. I have a long and abiding love for C.E.R.N.

>>RSM

ID: 19021 · Report as offensive     Reply Quote
Peter Ingham

Send message
Joined: 22 Sep 04
Posts: 6
Credit: 773,279
RAC: 0
Message 19179 - Posted: 7 Mar 2008, 8:07:16 UTC - in response to Message 19021.  

I'm seeing a few WU's that are reporting ~ 8% completed after 30 hours (Q6600) methinks something is badly wrong.


ID: 19179 · Report as offensive     Reply Quote

Message boards : Number crunching : To completion times for new tasks


©2021 CERN