Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 8
|
![]() |
Author |
|
aaabaaab
Cruncher Joined: Jul 31, 2016 Post Count: 42 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Hi everyone,
----------------------------------------I measured energy consumption (monitor excluded) for my pc (i7 4790 - 4 cores HT) and got the following results: # wcg tasks consumption (W) 0 29 1 55 2 71 3 85 4 97 5 98 6 102 7 108 8 111 It's running only Mapping Cancer Markers and it was configured to use the CPU at 80% so 6 wu were crunched at the same time. Average of WCG points: 50900/day. Two months ago I've changed the setting to 100% of CPU so 8 wu were crunched in parallel. Average of WCG points dropped to 42300/day. So energy efficiency dropped from 499points/W to 381points/W - a 23% drop. Is this making sense ? ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Did you set the % of CPUs to 80% or the % CPU time to 80%? The first would limit the nunber of cores used as you intended, the second would crunch on all cores but make short breaks to only let WCG run 80% of the time.
Assuming you did this right: the points per day of a specific project is not always the same and can change over time, as not all WUs are the same. Another possible explanation: thermal throttling could be a problem. What are your CPU temps and is your CPU runnung at the same clock speed in GHz in both cases? You could also check if the reported run time makes sense and is about (number of threads)*1 day per day on average over several days. If it is dramatically below that, there might be a problem. And yes - I think your thought is right that lower points output with more threads does not make sense. |
||
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7662 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Is this making sense ? No, it does not. MCM work units are very consistent in time and in points. However, the theory of thermal throttling does make sense. Make sure you have enough air flow and blow out the dust bunnies once in a while. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
aaabaaab
Cruncher Joined: Jul 31, 2016 Post Count: 42 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Did you set the % of CPUs to 80% or the % CPU time to 80%? 80% of CPU You could also check if the reported run time makes sense and is about (number of threads)*1 day per day on average over several days. If it is dramatically below that, there might be a problem. I've made an error in estimating the daily output. The PC is not running 24h/day so I computed the number of hours it was running from statistics here and then using the rule of 3 extrapolated to 24h. The mistake is that I've used the same formula for both periods: the one with 6 wu and the one with 8. After correction the results are the following: - with 6wu: 38000/day - with 8wu: 44000/day Thank you for pointing this out. ![]() |
||
|
plonk420
Cruncher Joined: Oct 18, 2006 Post Count: 18 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
hrm, my i7-4790 does 34-36K/day
----------------------------------------my 4c8t Ryzen 2400G does 37-43K/day for about the same watts INCLUDING GPU (undervolted by 0.1v), which is getting me ~65K BOINC points on MW@Home (possibly underperforming a little due to running only 1 stick of ram as it was so expensive when i got it) ![]() [Edit 1 times, last edit by plonk420 at Oct 15, 2018 10:55:47 AM] |
||
|
aaabaaab
Cruncher Joined: Jul 31, 2016 Post Count: 42 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
@plonk420
----------------------------------------Thanks for the info. Could you please let me know what type of wu is processing ? ![]() |
||
|
plonk420
Cruncher Joined: Oct 18, 2006 Post Count: 18 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
i want to say Smash Childhood Cancer
----------------------------------------![]() |
||
|
MrHasselblad
Cruncher Joined: Dec 20, 2014 Post Count: 42 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Take it a step up; although it works more effectively on a laptop.
Once a laptop is charged, and under most (but not all) settings - one can run that laptop while plugged in and then hooked up through solar/battery. It is surprising to see how small the panel is that one can get away with, also remembering that the battery reserve for the solar would still need to power the system through the change in seasons - and the possibility of a limited angle for the Sun. I'm running two systems this way; an Intel i9 laptop and also even though it is a desktop at 15oo watts - an Intel i9 18 core desktop. |
||
|
|
![]() |