| 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 |
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
My MCM WUs have dropped down to an average of 6-8 hours.
----------------------------------------CJSL Crunching for a better future... |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The mean runtime never got close to the [long] outliers as the below chart displays. The mini-charts at bottom now highlight peak-days and as can be seen we're already 4 days past the heaviest parts, which was 5.25 hours average for the day. Yesterday the project passed 400,000 validation for 1 day. Today the turkey is descending, Friday it is too, so not anticipating a matching number... about 375K results.
 |
||
|
|
hunter1978
Advanced Cruncher United States Joined: Apr 24, 2010 Post Count: 111 Status: Offline Project Badges: 
|
Not here. Getting some still running after 19 hrs. First time seeing this. Most tasks have been 3:30 to 5hrs. Intel i-7 3770k @ 3.50GHz.
---------------------------------------- [Edit 1 times, last edit by hunter1978 at Dec 1, 2013 2:06:15 AM] |
||
|
|
Eric_Kaiser
Veteran Cruncher Germany (Hessen) Joined: May 7, 2013 Post Count: 1047 Status: Offline Project Badges: 
|
Most tasks have a runtime around 3,5 hrs +-.
----------------------------------------But now I have some tasks in progress or finished (all starting with MCMI_0000216_*) with runtimes from 8 to ~20 hrs.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Started in a BOINC virtual machine [7.2.33] with a W7-64 environ [regrettably Linux was not offered during the setup], started 4 MCM simultaneous together with 4 CEP2. Efficiency as anticipated was less [1% or so], but what struck me was with batch 292, the MCM finished all within a 2minutes of each other, 3 waiting for a wingman:
7.26 mcm1 MCM1_0000292_0318_1 04:46:53 (04:41:42) 01-12-2013 14:07 01-12-2013 14:09 98.193 112.43 MB 74.57 MB 7.26 mcm1 MCM1_0000292_0547_0 04:44:57 (04:41:00) 01-12-2013 14:05 01-12-2013 14:06 98.614 112.09 MB 74.48 MB 7.26 mcm1 MCM1_0000292_0731_1 04:44:53 (04:39:33) 01-12-2013 14:05 01-12-2013 14:06 98.128 112.71 MB 74.93 MB 7.26 mcm1 MCM1_0000292_0693_0 04:44:38 (04:40:44) 01-12-2013 14:05 01-12-2013 14:06 98.630 112.43 MB 74.75 MB Memory usage... a very homogeneous lot. MCM1_ 0000292_ 0318_ 1-- Pending Validation 1-12-2013 14:12:00 1-12-2013 13:09:28 4,70 114,6 4,84 MCM1 1-12-2013 8:19:06 4,78 MCM1_ 0000292_ 0731_ 1-- Valid 1-12-2013 14:12:00 1-12-2013 13:06:20 4,66 113,5 88,5 4,79 18,99 MCM1 1-12-2013 8:19:06 4,75 MCM1_ 0000292_ 0547_ 0-- Pending Validation 1-12-2013 14:12:00 1-12-2013 13:06:20 4,68 113,5 4,79 MCM1 1-12-2013 8:19:06 4,75 MCM1_ 0000292_ 0693_ 0-- Pending Validation 1-12-2013 14:12:00 1-12-2013 13:06:20 4,68 113,4 4,79 MCM1 1-12-2013 8:19:06 4,74 At least, one valid, and this one with LAIM on and battery backup, so power outs wont generate restarts [not tried rebooting the VM yet to test how it recovers... that's the reason for the VM testing] |
||
|
|
rbotterb
Senior Cruncher United States Joined: Jul 21, 2005 Post Count: 401 Status: Offline Project Badges: 
|
Most of my MCM1 WUs are now running more in the 6-10 hr range. I still have my one WU in PV mode at nearly 35 hours. My wingman is due soon - I'm guessing it will be a no reply and another bigger cruncher will need to clear it out this week. But at least all my other MCM1 WUs that were in PV mode the past week or two have otherwise cleared out - so it looks like everyone is catching up on stuff and getting it done.
|
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
There is an issue we are investigating in which the client will show 100% complete but the workunit will continue to run. If this is the case you may want to consider aborting. We are looking into this issue. However if the client shows something less than 100% and you are seeing percent complete updating and CPU time being used I would recommend letting them complete. More than likely these are just long running workunits that have been sized improperly. As with any new research the researchers are still adjusting to running this application on a massive scale and runtimes should start to stabilize in the future. Since MCM1 has so many different types of workunits this may take longer than other projects.
Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My slow old machine was doing fine on these until they got REALLY long. Then, for some reason, it started claiming peanuts for points: MCM1_ 0000212_ 4739_ 1-- J-DESKSIDE Pending Validation 22/11/13 15:04:28 25/11/13 12:39:19 60.01 / 60.84 90.4 / 0.0 <-- ?!? You call that a slow old machine? Ha! This is a slow, old machine: MCM1_ 0000218_ 3872_ 1-- daffy Valid 11/22/13 18:48:35 11/30/13 19:18:01 154.68 / 154.68 443.2 / 345.7 Looks like things are returning to normal for me. Most work was in the 80-100 pt region originally, followed by a period when 300-400 pts was the norm. Last couple of days have seen a return of the 80-100 pt jobs. |
||
|
|
Eric_Kaiser
Veteran Cruncher Germany (Hessen) Joined: May 7, 2013 Post Count: 1047 Status: Offline Project Badges:
|
For me a longer runtime is not a big problem as long as the target date corresponds with the runtime. In my case my computer normally doesn't run 24/7 except during challenges or weekends. So a short due date and long (planned) runtime may be conflicting.
----------------------------------------But in the end: We are here to crunch the work away and bring this project a step further. |
||
|
|
|