| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: No CEP2 workunits on one of my PCs |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 14
|
|
| Author |
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges: 
|
Hello,
----------------------------------------Firstly, some background information: I have around 11 devices - from Athlon X2 to Intel Quad Q9550. What I observe, for several months, that my most powerful device (quad-core Intel Q9550, 4 GB RAM, plenty of free HDD space, fast Internet connection) virtually does not receive ANY CEP2 units. Other projects (DSFL, GFAM, C4CW) works fine and I receive WU from them without any problems. Other devices, with less capacity (even using the same Internet connection), receives CEP2 workunits quite regularly and I'm wondering why this single PC does not deserve the honour of calculation CEP2 workunits? In my device profile I have default setting regarding max. CEP2 workunits calculated in the same time (it means one). Any ideas?  [Edit 1 times, last edit by TKH at Dec 20, 2011 1:31:02 PM] |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Nope. My Laptop is set to 1 at all time and as soon as the CEP2 result is reported, very first work request after that fetches a new CEP2 [of course CEP2 box has been ticked :o]. This science is kind of in the same work fetch class as repairs... they get pushed out by priority before anything else, if a slot is free.
----------------------------------------Why not getting them? Have a look in the message/event log after setting CEP2 as the only project [of course considering that the default profile is the one that is associated to the device]. If something is not suitable to the server, it will be printed. If there's anything 'duh' in the above, sorry, but we have to consider all possibilities of oversight. --//-- P.S. Your background info omits OS info. [Edit 1 times, last edit by Former Member at Dec 13, 2011 3:18:25 PM] |
||
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges:
|
OK, so I created a separate profile only for this single PC, checked only CEP and forced project update and CEP workunit was immediately downloaded without any warnings or errors.
----------------------------------------I believe that I read somewhere on the forum that WCG servers evaluate each machine according to its computational power, capacity, RAM size etc. and then sets different project priorities. If this is true, maybe this machine was evaluated in the way that other projects have higher priority than CEP? I have no idea :( EDIT: OS is Windows Vista 32bit, BOINC client is 6.10.58 (current release, downloaded from WCG). [Edit 1 times, last edit by katoda at Dec 13, 2011 5:56:53 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Suggest you visit the profile the client was attached to, untick CEP2, Save the profile and, then reopen and tick CEP2 again and save. Then hit update in the client and compare the memory / disk settings, which are the main controls that are evaluated on each work request.
If any of the settings were too low and you had temporarily selected only CEP2 in the original profile, you would have had a confirmation in the client message log if the settings are sufficient on that profile. Maybe compare what was printed in log what the mem/disk settings were of the old and the new profile. The bit to compare is this: 186 World Community Grid 13-12-2011 19:03 General prefs: from World Community Grid (last modified 12-Dec-2011 17:16:01) 187 World Community Grid 13-12-2011 19:03 Computer location: school 188 13-12-2011 19:03 General prefs: using separate prefs for school 189 13-12-2011 19:03 Reading preferences override file 190 13-12-2011 19:03 Preferences: 191 13-12-2011 19:03 max memory usage when active: 2067.17MB 192 13-12-2011 19:03 max memory usage when idle: 2657.79MB 193 13-12-2011 19:03 max disk usage: 28.32GB 194 13-12-2011 19:03 don't use GPU while active 195 13-12-2011 19:03 suspend work if non-BOINC CPU load exceeds 45 % 196 13-12-2011 19:03 (to change preferences, visit the web site of an attached project, or select Preferences in the Manager) As for computational power, think there's a limit somewhere that no CPU gets work below 256Mhz. The client itself computes if it can complete work in time [not enough hours on per day]. That too would get printed in the message log. Else, cant think of a reason. My clients all get CEP2 first. --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello katoda,
How big is your cache? If you only get 1 CEP2 work unit, it will wait until all the earlier work units are run, then it will run, then after it is returned you can draw another CEP2 unit. If you have a 2-day cache, you can only get 1 CEP2 unit in 2 days. If you want a lot of CEP2, keep a very short cache. My extra cache is 0 and I keep a basic cache of 0.1 days (2.4 hours). So every day my quad runs 1 CEP2 unit and starts on a 2nd. Very roughly I run 1.5 CEP2 units a day. Lawrence |
||
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges:
|
OK, here you have what is posted in log of BOINC agent regarding mem usage:
----------------------------------------2011-12-14 19:21:31 Reading preferences override file 2011-12-14 19:21:31 Preferences: 2011-12-14 19:21:31 max memory usage when active: 2494.59MB 2011-12-14 19:21:31 max memory usage when idle: 2993.51MB 2011-12-14 19:21:31 max disk usage: 4.00GB 2011-12-14 19:21:31 don't use GPU while active Regarding cache, the device profile have it's default setting - "Cache 0 days of extra work". Just for clarification, this machine uses the profile together with around 6 other PCs and the other machines receives CEP2 workunits without any problems, so I doubt if the problem is caused by profile settings. I will temporarily increase the max number of CEP2 workunits and see it anything will improve. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You'll absolute find out if you temporarily change the shared profile to only assign CEP2 work. Briefly suspend the work fetch on all your other hosts, then increase the additional buffer a little to force a work fetch.
Also, your disk allowance is set to 4GB, but how much is already used and what percent is set to leave free? [4GB, with a 50% free means only 2GB is allowed]. You can see in the BOINC Manager Statistics tab how much is already in use. You wrote yesterday that a separate profile for this host immediate send a CEP2 task, so do not think space is the answer. Still, I'd suggest you open local preferences and hit the Clear button, to be sure that only website profile settings are in effect. Do that after the trial of increasing the buffer in the local prefs. --//-- |
||
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges:
|
Well, some more information about what I found yesterday.
----------------------------------------As I mentioned before, after assigning this machine to a separate profile with only CEP2 checked, I downloaded CEP2 workunit immediately. Then I reassigned back this machine to it's default profile, but then I increased number of CEP2 workunits from 1 to 2. Then this machine downloaded 2 CEP2 workunits - again without any problems nor error messages. Now I changed this setting back to the default (1 workunit) and have to wait a little to see, what will happen when both CEP2 workunits will be returned. Regarding disk space usage, the limit is 4 GB, as you can see, and according to BOINC agent, current usage is around 800 - 900 MB. When a CEP2 workunit is processed, it increases to around 1,4 GB. There is around 70 GB of free space on the partition where BOINC data folder is located, so the setting "leave at least 2 GB of free space" is not violated :) |
||
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges:
|
So both CEP2 units were returned, local setting were cleared (the only difference is that for this machine I've set 80% limit on CPU usage... and again, no CEP2 workunits. I give up.
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
May I suggest that you switch all your devices to the test profile you had where 1 would get to that host. Then after, delete the resisting profile (but it's named ''default'', don't think you can). Actually I don't use default, only school, work, home... default could be anywhere :D
Beyond that "Resistance is futile, we're the Borg." --//-- |
||
|
|
|