| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Why is the BOINC Elapsed Time different from Status Results CPU Time? [RESOLVED] |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
OK, I have to ask ... like the title says, why the difference... I'll try to give an example:
----------------------------------------When a WU goes complete, let's say that the ""Elapsed Time" that the BOINC manager reports is 3.06 (3hr, 6 mins). When the WU shows up on the Status Results page, it's showing something like 2.69 (2 hr, 41mins). I'm sure there's a logical/technical reason for this, I'm just curious what it is. I tried the FAQs, but didn't find anything that answered my doubt. Thanks, CJSL ---------------------------------------- [Edit 1 times, last edit by cjslman at Jan 8, 2012 12:20:31 PM] |
|||
|
||||
|
KWSN - A Shrubbery
Master Cruncher Joined: Jan 8, 2006 Post Count: 1585 Status: Offline |
In your example elapsed time is 3.06 while CPU time is 2.69. Elapsed time is based on wall clock where the CPU time is calculated based on efficiency. If the CPU is waiting for data to crunch, time isn't being counted. If you're running other programs that take CPU cycles away from WCG, that time isn't counted.
----------------------------------------Therefore the stats report shows the actual amount of time that a given CPU was devoted to WCG rather than the actual time it took for the work unit to finish.  Distributed computing volunteer since September 27, 2000 |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges:
|
Thanks for the response... OK... that sounds reasonable. I thought that the elapsed time stopped, when it wasn't crunching, but I guess I that's not always the case. Now, does this also have to do with the "% CPU Time" that I have specified in the Preferences? For example... if I specify 85% CPU Time, the difference between Elapsed Time and CPU Time would be aprox 85%?
----------------------------------------Thanks, CJSL |
||
|
|
KWSN - A Shrubbery
Master Cruncher Joined: Jan 8, 2006 Post Count: 1585 Status: Offline |
Someone else is going to have to comment on the preferences as that is beyond me.
----------------------------------------Elapsed time does stop when it's not crunching. If it crunches for 10 minutes at 90% efficiency (10% going to other programs) then you'll see 10 minutes elapsed 9 minutes CPU as an example. Distributed computing volunteer since September 27, 2000 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Whilst cjslman says he could not find anything in the FAQ's on this topic, it so happens that there is expanded text in 3 posts under the second item in the Start Here FAQ index. Ctrl-F then type Elapsed and the browser page jumps to the key word topic.
The percent of 85 in the example, would be an "at best". The system even if left untouched does all sorts of stuff. The more apps open and system services running, the bigger the differential. Here's where the die hard tuners will research what can be switched off... my Linux file/printer server [no GUI loaded] manages 99.85% on Clean Water for instance. It really does not have to do much else. But, when using the system, whilst BoINC is set to use 100% of *spare* cycles, the BOINC efficiency can still drop to 85-90% of total elapsed time, speak CPU time reported shows a large gap. This is most pronounced for this system on CEP2. --//-- |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges:
|
Sekerob... thanks for the pointer to the FAQs that explains the difference between elapsed time and CPU time (not sure how I missed it
---------------------------------------- ). And now I think I have a handle on that difference. As for how the "% CPU Time" setting fits into this, I think I understand that also (I won't try to muddy the waters of this thread going into that). What I do want to do is thank KWSN & Sekerob's effort in explaining this to me (and others who see this thread). Happy Crunchings !!! CJSL |
||
|
|
|