| 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: Only get 1 workunit on a 8 Core machine, no ambition to crunch anymore!!! |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hallo,
----------------------------------------I crunch from time to time for world community grid. I setup a virtual machine with 6 GB RAM and 8 Cores and what do I get? Only 1 workunit! Whats that **? When I request more work the only reply is that there is no work anymore. **, there is definitively enough work, otherwise the other machine (see below) wouldn't have gotten work. I setup another machine, 4 Cores, 4 GB RAM and also got only 1 workunit. I changed settings so that all cores are used all the time. When I can't use 100 % CPU time because I don't get enough work units I don't crunch anymore. That's very said. I think you make wcg for other users also unattractive when you don't give them as much work units as they need. I have a grafic card with CUDA but I can't use it, because I set up VMs and thats because your BOINC client spys and tells WCG details like MAC-Adress, hostname and local IP-Adress. I'm off until I get 8 workunits and 4 work units for my other machine I won't contribute computing power to WCG anymore! **Edited for appropriate forum language** TKH [Edit 1 times, last edit by TKH at Oct 31, 2011 1:21:46 PM] |
||
|
|
krakatuk
Advanced Cruncher Germany Joined: Oct 3, 2008 Post Count: 141 Status: Offline Project Badges: 
|
Did you select only Clean Energy Project? In this case it would be pretty clear, why you get only one WU a time... This project has some default restrictions.
----------------------------------------To fix that you have to go to your device profile settings (try here: http://www.worldcommunitygrid.org/ms/device/v...iguration.do?name=Default ) Then find "Project Specific Settings" and select "Unlimited". It should help.  |
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2955 Status: Offline Project Badges: 
|
Also, further to what krakatuk has suggested, it is a recommendation to select the "If there is no work available for my computer for the projects I have selected above, please send me work from another project." option - as, that way, you'll virtually be guaranteed to get work.
----------------------------------------PS. Along with the specific requirements for CEP2 (did you check them out if that's the project you're crunching? - there is a reason why that project is an "opt-in" science), currently DDDT2 hasn't any work at all to process and I do believe C4CW has very little work. Added to that, HCMD2 has been 'off stream' for a little while, and should be back up and running soon...  [Edit 1 times, last edit by gb009761 at Oct 31, 2011 11:05:16 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
8 or 4, speak all cores in a VM (and probably HT for the 8 core at that). Strongly *not recommended* lest you've tried and tested that the storage IO can handle this. I've found my quad to do well with 2, starts sweating at 3 (increased gap between CPU and walltime) and on 4 looses ability to go straight up. Choose something light on the side such as the worthy Clean Water or Help Conquer Cancer (HCMD2 is lightest but unavailable presently).
Oh and at high use (3 or 4), W7 far outperforms (bare) Linux on my quad for CEP2. --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
To fix that you have to go to your device profile settings (try here: http://www.worldcommunitygrid.org/ms/device/v...iguration.do?name=Default ) Then find "Project Specific Settings" and select "Unlimited". It should help. Thanks that was it. Now I get enough work units. Why is there such a silly default setting? It would be great if in the messages window there would be written "no work available - please check default settings in your profile" 8 or 4, speak all cores in a VM (and probably HT for the 8 core at that). Strongly *not recommended* lest you've tried and tested that the storage IO can handle this. A HT core is better than no core, of course it is not as fast as a real core, but surely better than no one. Why should there be storage problems? Setting is to write once per minute to disk. This works fine, I have no I/O problems but 4 and 8 cores working :). |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The emphasis is on *IO*, and if it fine for you, that's great, but do watch the CPU to Elapsed time ratio, which is where the sneak creeps in. You wont notice but multiple tasks start writing to storage at same time, there could be major bottle necking. Normally, projects at WCG can run 99% and better efficient. With CEP2 loaded on all cores, that is for many not the case... jerky mouse, screen refresh lag a sign that the crunch is on.
Settings set to 5 or 10 minutes would improve performance, but not for CEP2. It will just ignore that setting and write it's checkpoints when a job is ready. BOINC was originally never designed to handle more than 1 project per grid., so what can be messaged on a per-science basis is limited. That said, CEP2 IS opt in and there is a clear warning on the device profile to read the RTFM, or did you miss that piece? --//-- P.S. Strom_umsonst or Storm_umsonst? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Dear Strom_umsonst,
For more information please also check out the recommended settings for CEP2 in the footer of this post - it also contains a few explanations about things. Best wishes Your Harvard CEP team |
||
|
|
|