| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Impact of the Benchmark |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 13
|
|
| Author |
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: 
|
Setting my buffer to 10 days, I get 13 WU that my system thinks is going to fill my buffer for 10 DAYS (it has actually forced my system into high priority) and it will in fact be empty in less than 12 HOURS. Not happy I am lost a little in this reasoning, or I miss something. I have a 1 day buffer that indeed sometimes has a little number of WU's when they have high expected CPU time values. But I have also a permanent open connection to WCG, which means that there are multiple times over 24 hours were an automatic update is made. This means that the buffer in reality never empties. I never saw one machine idle because of an empty buffer. This happened only when the wireless link was cut for various reasons and the device could not communicate anymore. When I change projects through profile changes, I do systematically manual updates, rerun the local configs run the benchmarks etc.., to avoid any memory effect. If you connect only once a day then yes you may have a problem.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I am lost a little in this reasoning, or I miss something. The point is that I set my buffer to be what I wanted (using 10 days as an example) and instead of getting what I requested, I get 5% of what I wanted (10 days / .5 days). How does the system expect me to be efficiently processing WU when the fpop estimate on a WU is out by 2000% ? |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
astrolab, I got your point. Yes indeed there will be an efficiency rating that will not be on paar with reality.
----------------------------------------I never asked myself this question, and I wonder how are my devices rated. But I remain a little primitive on that. As long as my buffers do not empty and machines crunch without being interrupted, AND I get no Server Aborts all is fine for me.  |
||
|
|
|