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: 566
|
![]() |
Author |
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
1200 aborted, 4691 aborted [more that 1/3rd of the 11691 my comps did to complete the 2 years]. Well, I've already passed 22k results, and would probably land close to 25k for 2 years. On the particular computer, 1200 CFSW-tasks equals roughly 4.4 days cache-size. ![]() "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I think if the most of the top crunchers are abandoning loads of WUs as it seems happening, then i guess the left 10 days should be doubeling :) Well, I'm not really a top cruncher but around 14 days ago I enlarged my caches up to 7 days. Now I see two of my crunchers canceling some WUs because they will not match the deadline. Right now I have around 35 abandoned WUs and there will be more ... I guess another 40 WUs. Don't ask me why this happens ... the systems were running on this project long enough and the boinc should we able to guess the right numbers of WUs. But something went wrong ... and I have no idea. Those guys with larger caches will release more WUs if they got hit by the same mystery. So I would not be surprised if there will be a lot of WUs coming back from overloaded(tm) caches. May be there is enough work for 20 days ... but who knows. In the meantime I have reduced my caches to 4 days and I think I can reach for sapphire in 10 days. I will reduce my caches when I think I have enough reserve WUs to get this badge. Stephan About 4 days ago, I increased my cache to 6 days on two computers to obtain enough WUs to get me through to two years. After the caches finished loading, I changed my cache to one day so I wouldn't, unnecessarily, get more WUs than I needed to get to two years. I do not have any error or aborted WUs. I also do not have any abandoned WUs as near as I can tell. At this time I have 1 year 325 days so I expect to reach 2 years in the next 4 days. |
||
|
RaymondFO
Veteran Cruncher USA Joined: Nov 30, 2004 Post Count: 561 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The real problem that I see is that certain individuals will "over-cache," and then release WU's they will no longer crunch since and they did not budget the amount of WU's they will need to reach a given desired result. This may result in a certain disruption for WCG processing WU's, but then again I have been a beneficiary of individuals who are considerate enough to release WU's so others may reach their project crunching goal. Going back to what I believe was stated a elsewhere, cache sizes should be limited to a per core basis and anything approaching ten (10) day cache is in my opinion excessive.
----------------------------------------My own cache size used to be 0.00 (except for one machine at 1.25 days) however that was changed to 0.5-0.75 days (again, except for one machine at 1.25 days) because the recent WCG reporting results issue, that now appears to have been resolved. Based upon my own forecasting and rate of completing WU's, I anticipate reaching my goal before the project is completed. By simply using appropriate allocation and budgeting of WU's the need of having large and excessive cache size is unnecessary. [Edit 1 times, last edit by RaymondFO at Sep 11, 2012 5:26:37 PM] |
||
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
1200 aborted, 4691 aborted [more that 1/3rd of the 11691 my comps did to complete the 2 years]. Certainly, if there's a cap of 80 or 120 per core per day, there might be moments needed to consider capping total assignments per device [ knreed is reading this maybe] to stop runaway caching. How will that work with HCC GPU at 6 minutes per run [well the multipacks WU's would then be 30+ minutes if e.g. it would be 5 HCC units in a WU]. If not the cap would have to be upped further with all the associated further risk of cache inflation. The limits in force are 30 results per scheduler request and up to 150 results per day per core. At the moment this allows those who desire caches, to build them up. I apologize if I missed it, but are there people who are incorrectly building up caches? What I understood above is that people have large caches, but as they reach their goal they are aborting them to return them for others to run. I'm not seeing any impact to distribution so on the server side, I'm not seeing an issue at this time. However, please let me know if people are getting caches they didn't ask for built up. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It was this comment
Back to CFSW, on my severely over-cached computer just aborted roughly 1200 tasks, since no hope of finishing these before their deadline. and another mentioning the 4691 aborted by hand that made me wonder if overall per-device cache capping may be called for. |
||
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Current estimates show a little over a week left of new work to be sent out.
Thanks, armstrdj |
||
|
pcwr
Ace Cruncher England Joined: Sep 17, 2005 Post Count: 10903 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Current estimates show a little over a week left of new work to be sent out. Thanks, armstrdj ![]() I currently need 21 calendar days to get to Emerald. Patrick ![]() |
||
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 172 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I need 13 days, daily increase of statistics is about 5-6 days, so I will make it :)
----------------------------------------![]() |
||
|
NiceMedTexMD
Veteran Cruncher United States Joined: Aug 17, 2006 Post Count: 929 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I need about 12 days. Have to at least try. :) Will be interesting too in the future what the data shows and is used for.
----------------------------------------Dr. Mike ![]() |
||
|
rilian
Veteran Cruncher Ukraine - we rule! Joined: Jun 17, 2007 Post Count: 1460 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
NiceMedTexMD - easy. pack 10 days of cache on 14th september :)
---------------------------------------- |
||
|
|
![]() |