| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: Neverending MCM tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 86
|
|
| Author |
|
|
MJH333
Senior Cruncher England Joined: Apr 3, 2021 Post Count: 300 Status: Offline Project Badges: 
|
Sgt. Joe, Grumpy Swede, alanb1951 and Unixchick,
Thank you for the interesting discussion about the differences between the LOO and NFCV tasks. I've just done a not-very-scientific eyeballing of my recent results. On one of my Ryzen processors running on Linux, the NFCV tasks seem to take roughly 25% longer than LOO. On an identical Ryzen processor running on Windows, the LOO tasks seem to take roughly 10% longer. I also run MCM on Android on an Odroid N2+ and some Pi 4Bs. The NFCV tasks take roughly 60% longer than the LOO tasks on the Odroid. On the Pis, the NFCV tasks take nearly 5 times (!) as long as the LOO tasks. Weirdly, the Pis are quicker than the Odroid on the LOO tasks. I have no technical insight into any of this stuff, but I thought you might be interested in these results. I find the Pi vs Odroid results baffling! Cheers, Mark |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2495 Status: Recently Active Project Badges: 
|
Time to add to what I said about NFCV and LOO, when it comes to CPU heat.
----------------------------------------I said that NFCV runs hotter than LOO on my Windows/Intel computers, and that was true when it came to My Laptops. One with a i3 380M CPU (Windows 7), and the other with a i7-3630QM CPU (Windows 8.1). When testing now with my desktop with a i7-4790K CPU (Windows 8.1), that statement is being reversed when it comes to that computer. The LOO runs considerably hotter than NFCV on my i7-4790K computer. However,what I said about NFCV being faster than LOO, still stands when it comes to all three computers. [Edit 1 times, last edit by Grumpy Swede at Nov 21, 2021 1:08:54 PM] |
||
|
|
Felix Kaeufer
Cruncher Joined: Feb 3, 2012 Post Count: 29 Status: Offline Project Badges: 
|
Kevin,
Is there a fix for the pause/restart issue? Maybe, something I can do on my part? Cheers |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Most computers do not experience the issue of long running mapping cancer markers jobs. However, for those that do, the only workaround is to set your preferences to use 100% of processor time.
|
||
|
|
Felix Kaeufer
Cruncher Joined: Feb 3, 2012 Post Count: 29 Status: Offline Project Badges:
|
OK. I'll simply test from time to time, if the issue is gone.
|
||
|
|
psmith95
Cruncher Joined: Sep 14, 2021 Post Count: 6 Status: Offline |
The work around seems to work. I set the CPU time to 100% and restarted sone of the tasks that I had suspended. They are all now past 15 minutes and seem to be continuing to run. I am on macOS 12.0.1,Monterey running on INTEL I7.
This problem has been reported over a year ago. Is anything being done to correct it? Is it a problem with the WCG vesion of Boinc manager, Boinc manager in general, or is it with the MCM application, since that seems to be the only one affected |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
The work around seems to work. I set the CPU time to 100% and restarted sone of the tasks that I had suspended. They are all now past 15 minutes and seem to be continuing to run. I am on macOS 12.0.1,Monterey running on INTEL I7. This problem has been reported over a year ago. Is anything being done to correct it? Is it a problem with the WCG vesion of Boinc manager, Boinc manager in general, or is it with the MCM application, since that seems to be the only one affected The problem isn't that old, it was reported on the forums a couple of months ago. I don't what the issue is but it is limited to the MCM app and on MacOS.  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
psmith95
Cruncher Joined: Sep 14, 2021 Post Count: 6 Status: Offline |
It now appears that the work around did not fix the problem, just changed it. The tasks are running and elapsed time is increasing as it should. However, I have two MCM tasks estimated at 3:31:35 that are still running after 20 and 16 hours.The estimated time to completion is reducing very slowly, much slower that the elapsed time is increasing. For one of my tasks, it took over a minute to reduce the estimated time remaining by one second. It definitely seems that something is wrong with the MCM application.
I will try to finish the remaining MCM tasks that I have, but I think I will be dropping MCM until this is corrected. This is delaying real work on other WCG tasks and tasks for other projects in which I participate. I would rather have my computer doing real work. |
||
|
|
psmith95
Cruncher Joined: Sep 14, 2021 Post Count: 6 Status: Offline |
Falconet is correct. I think I noticed the joined date of a submitter since it is bold near the top of the message instead of the message date which is in normal text at the bottom of the message.
|
||
|
|
Nick Batos
Cruncher Joined: Apr 5, 2020 Post Count: 7 Status: Offline Project Badges:
|
After a number of attempts, I was able to get MCM to occasionally complete using only 88%CPU time. Anything lower and the MCM work units would always restart after 2 to 3 minutes. Using 100%CPU time has allowed me to complete MCMs consecutively. So I was able to use this 'workaround', but obviously this isn't a 'fix'. [10 core M1Pro MacBook Pro, macOX v12.0.1, BOINC v7.16.19]. Given that I don't want to run at 100%, I too will stop MCM until this is fixed, too bad...
|
||
|
|
|