| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: How total run time is measured? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi !
Two questions from newbie: If i have 6 core processor and boinc using 100% then my total run time is 6x real time ? Is there any comparison of cpu performance in boinc projects? Sorry for my bad english :) |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
All there is to know about time accounting on My Contributions from 'Help ': https://secure.worldcommunitygrid.org/help/viewSearch.do?searchString=elapsed
|
||
|
|
twilyth
Master Cruncher US Joined: Mar 30, 2007 Post Count: 2130 Status: Offline Project Badges: 
|
The part about GPUs I assume is new since HCC.
---------------------------------------- "Run time" is simply a measurement of how long a volunteer's device spends running World Community Grid tasks. You accumulate run time for running tasks on your device's GPU, CPU, or both. If you contribute GPU time, the run time is the actual elapsed time in which the World Community Grid task executes on the GPU. In other words, if the task starts at 10:15am and finishes at 10:19am, you accumulate 4 minutes of run time. The reason I say that is that I only accumulated about 40 years of run time for HCC but did 1.4M results. For comparison, I have 100 years for FAAH and only 350k results. Also I have 50 years for MCM but less than 100k results. So obviously there is variation between projects but there is a huge disparity between HCC and other projects in terms of the amount of run time that was credited. Also I have this vague recollection from when HCC was running that time was only being credited for the cpu computations.   |
||
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges: 
|
Hi ! Two questions from newbie: If i have 6 core processor and boinc using 100% then my total run time is 6x real time ? Is there any comparison of cpu performance in boinc projects? Sorry for my bad english :) Actually it's almost 100%...but to be sure: 1. go in BOINC DATAdir 2. open file named sched_request_www.worldcommunitygrid.org.xml DO NOT CHANGE ANYTHING INSIDE THE FILE! 3. search for <time_stats> 4. few rows don't will be listed <active_frac> That is the availability of your CPU for BOINC!  mine says: <active_frac>0.999889</active_frac> which means 99,9889% my CPU is crunching data on WCG. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The part about GPUs I assume is new since HCC. ... The reason I say that is that I only accumulated about 40 years of run time for HCC but did 1.4M results. For comparison, I have 100 years for FAAH and only 350k results. Also I have 50 years for MCM but less than 100k results. So obviously there is variation between projects but there is a huge disparity between HCC and other projects in terms of the amount of run time that was credited. Also I have this vague recollection from when HCC was running that time was only being credited for the cpu computations. Blond moment, if HCC-GPGPU tasks, identical to the CPU ones, and run 18-20 times faster than on CPU, then how much time [in relative terms] do you expect to get logged for 1.4M results? And the HCC-GPGPU elapsed [clock] time was logged for the GPGPU tasks, just as the FAQ says. IIRC, it was even a hold-up before launch, to get that coded right. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi ! Two questions from newbie: If i have 6 core processor and boinc using 100% then my total run time is 6x real time ? Is there any comparison of cpu performance in boinc projects? Sorry for my bad english :) Actually it's almost 100%...but to be sure: 1. go in BOINC DATAdir 2. open file named sched_request_www.worldcommunitygrid.org.xml DO NOT CHANGE ANYTHING INSIDE THE FILE! 3. search for <time_stats> 4. few rows don't will be listed <active_frac> That is the availability of your CPU for BOINC! mine says: <active_frac>0.999889</active_frac> which means 99,9889% my CPU is crunching data on WCG. Actually -Not-. You can have 0.600000 for <on_frac> and 0.999999 for <active_frac> and only compute 24*.6*.999999 hours a day, the equation used to proportionally buffer work. E.g if you set 1 day buffer, but the device is -On- only 0.6 part of the day on average [last 10 days or so], only 6/10th of 24 hours will be buffered. Bar that, BOINC only uses 'idle', spare cycles, so the daily runtime is process threads allowed*24 less system time less user app time less efficiency differential [when set to run always]. <active_frac> can thus be substantially misleading to infer hours per day recorded. You may edit anything you like in sched_request_www.worldcommunitygrid.org.xml. The file is refreshed with each new server connect from data in the client_state.xml. |
||
|
|
|