| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Website Support
Thread: Average turnaround time |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 16
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi, I've searched everywhere in 'my Grid', but I couldn't find the average turnaround time for my machine.
Is there even such a value at wcg? Or don't you just display it, because I'm pretty sure it will be used as measurement for betas. Thanks in advance and Cheers -Wiesi |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Wiesi,
My Grid lists Avg. Run Time Per Result. THere is no listing of average turn around time. The staff see a lot of statistics about the grid that are not published, but I doubt they try to break things down to individual members. Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Has not been published for a long time but over 90% in past comes back within 4 days from issue. It's a gradient in fact, where day 4 is lots less than day 3, most in fact coming back in the first 48 hours.
----------------------------------------There was a post by knreed 2-3 years ago. Don't expect that to have changed much. [Added] The servers do track the return times of devices to determine if under / over 21 hours to put them in the "reliable" bin.[/Added]
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 2 times, last edit by Sekerob at Jan 22, 2009 2:52:01 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The servers do track the return times of devices to determine if under / over 21 hours to put them in the "reliable" bin. Does that mean that if a computer is running a 2 day buffer, it is automatically not 'reliable'? |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
We use this term for lack of a better word, the function being to identify the devices that "reliably" qualify to be considered for rush jobs to get quorums and batches completed quickly.
----------------------------------------A device that is returning 15 results consecutively is also consider "reliable" in terms of work requests, regardless of the time it takes to return a result. It gets all it wants up to the limit of 10 days buffer restriction (sum of all active projects attached to a device), but wont be send rush tasks, para 1. This was written up in the FAQs in different words. [edit]appended quickly to first line.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Jan 22, 2009 8:29:47 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi, I've searched everywhere in 'my Grid', but I couldn't find the average turnaround time for my machine. Is there even such a value at wcg? Or don't you just display it, because I'm pretty sure it will be used as measurement for betas. Thanks in advance and Cheers -Wiesi Yeah, I've looked at this in the past and wondered too... but... not been able to find any data that is avaialble for you n' me to peruse.. Now, if your interested in your own system, here's what you can do... change your settings to look like this:  what it basically does is to check about every n units of time to see if there's anything to upload... and down load.. Of course, this takes time away from processing... somewhere I put forward the notion of having the program (it knows how much longer it has to run until it's complete) start poling the network about a minute from completion so your job gets as much cpu resourse as possible... now, if you get say 10 days of WU ... it's kind of pointless. tradeoff's in all directions. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello BarneyBadass,
Connect 10,000 times a day! Isn't that overdoing it? Connecting a hundred times a day will connect every 15 minutes. I cannot imagine anything gained by doing it faster. Going below 0.01 is just silly. (I think.) Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Not sure what BBA's post does other than telling how to keep the turnaround time to the lowest possible. Once there is work in the queue of 0.1 days or more, including work in progress, the client wont contact the servers until the next work call, to same time clear the "Ready to Report" line.
----------------------------------------Notably, BOINC will upload result files whenever the line is open regardless the connect setting. Not even if "confirm connect" is set. That works only on dial up., but i could be wrong. Gain on the Reliability scale: Zero. But, never too old to learn new tricks, and this fox has not lost it's gray hair.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Sekerob and lawrencehardin,
I agree completely with you, however the OP's question was: Hi, I've searched everywhere in 'my Grid', but I couldn't find the average turnaround time for my machine. Is there even such a value at wcg? Or don't you just display it, because I'm pretty sure it will be used as measurement for betas. I posted a means by which the OP could get the average turnaround time, because it tends to start to pull the next wu just before the curent one finishes and this process is repeated over and over and over. So, that being said, that's what it does... its overkill, but it's the best that can be done without some code modifications. and I can't imagine the amount of time in the polling sequence of about 10,000 x a day is all that consumptive but I'm not a windows dispatching jock so I dunno. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
I may be wrong, and Didactylos will probably tell us what is the correct answer, but my understanding of the "Connect about every..." parameter is that it is not at all an instruction to connect for Boinc, it is only an indication you give it, for it to compute how much work it must keep buffered to ensure continuity of crunching.
----------------------------------------Remember, in older versions that was the only parameter for that purpose, and if you play with both parameters (this one and the "extra work" one) you will find that it is the total of both that matters for queuing WUs. Boinc connects only when it needs to, or when you force an Update. Cheers. Jean. |
||
|
|
|