| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Strange "mass download" of DDDT2 tasks to my machine |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 4
|
|
| Author |
|
|
JollyJimmy
Advanced Cruncher USA Joined: Aug 23, 2005 Post Count: 115 Status: Offline Project Badges: 
|
Given the (historically) rather limited availability of DDDT2 tasks, a few days ago one of my machines experienced the equivalent of "rain seeker's paradise" - sort of: For whatever reason, and in spite of a 2.5 day cache, that machine, an eight core 64bit Wintel, managed to download literally HUNDREDs of DDDT2 tasks.
----------------------------------------Took me a while to realize that, and also took me a while to overcome the resulting euphoria, only to realize that there is no way my machine will be able to finish even a small percentage of that work within the assigned time frame. So obviously something went seriously wrong there and should probably be investigated. Here are some additional details: - I have a second machine, virtually identical (bit less RAM or so), that did not get to go to temporary paradise, even though on the exact same profile. - I aborted somewhere about 450 tasks that would otherwise be due on 03/19-20. - There is still a boat load of other tasks, due at a later time, that I plan to abort later this week (unless you tell me otherwise). What's going on?  |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You must have been running on empty and then connected right at the time when a wave of results expired... due the 19-20th means most probably at that quantity "No Reply" make up jobs.
As for the times, the DDDT2 C Type work as like 8 sub types, the 's' the short, taking about 1.5 hours here, so an eight core could well be completing them in 4 days. The 'p' range here from 5 to 6 hours. Why so many? You know best, since you can look at the host DCF. Probably it was shot after maybe a series of short HCMD2. They vary wildly from 3 minutes to 10 hours, the range my quad has seen in the past week. Enjoy your luck and crunch on. --//-- |
||
|
|
JollyJimmy
Advanced Cruncher USA Joined: Aug 23, 2005 Post Count: 115 Status: Offline Project Badges:
|
Not really, Sekerob:
----------------------------------------I have been running CEP2 and DDDT2 almost exclusively for months, with only a few betas and HCCs sprinkled in. Nothing else. I believe at the time this thing happened, I had nothing but DDDT2 in the queue. The queue was likely full or at least well filled at the time (I have also recently reduced my cache size; did that before this happened, but not sure how close it actually was), and I have not observed any "ultra short" DDDT2. Anyway, if you are not concerned, then I am not concerned, so I will be a happy cruncher  .I just hope my reliability rating will not get dinged too much by this amount of user aborts - still hoping to get a Gold in Beta one day...  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
450 task aborts are kind of like dinged for a good bit ;>)
----------------------------------------(it will have been out anyhow with a 2.5 day cache since a ''reliable'' device must return results consistently in less than 2 days, except now you're way at bottom of ladder, so the device will have to return at least 77 valid results without interruption **. Peanuts for an 8 core ;P) --//-- ** 77 from a clean start, and lots more if totally messed up. edit: Yes it seems it can run up to 122 valid results consecutively before a device is fully reliable again. [Edit 2 times, last edit by Former Member at Mar 16, 2011 7:06:37 PM] |
||
|
|
|