Task Overload
Message boards :
Problems and bug reports :
Task Overload
Message board moderation
Author | Message |
---|---|
Send message Joined: 19 Nov 14 Posts: 93 Credit: 30,066,240 RAC: 0 |
Hi, I've had some hardware problems, now hopefully sorted out. When I got my rig running again, since I'd lost all the data in the projects directory I reset a@h. Now I find I have a large quantity of WU that each run for more than 3 hours, there is absolutely no way my rig can complete all the WU. I cannot run 24/7 only about 7 hours per day.. So what do I do about the WU on hand? I don't want to abort a shed load of WU if I can avoid it. Regards, Cliff |
Send message Joined: 19 Jun 12 Posts: 20 Credit: 18,088,949 RAC: 2,577 |
|
Send message Joined: 1 Jan 14 Posts: 300 Credit: 32,264,344 RAC: 7,374 |
Hi, Lower your cache size first or you will just get them all back again! |
Send message Joined: 25 Jul 14 Posts: 64 Credit: 100,582,080 RAC: 0 |
Lower your cache size first or you will just get them all back again! Hi! Is it really possible the same host would get them for the second time? I don't know anything about this... but that information started to tickle my mind. I thought that kind of situation wouldn't be somehow logical behaviour by the project server. |
Send message Joined: 19 Nov 14 Posts: 93 Credit: 30,066,240 RAC: 0 |
Last modified: 7 Jan 2015, 14:37:44 UTC Hi Mikey, Its well lowered, but in any event I only used to get 4 WU at a time, and they were all sse3 type, now I have 1 sse3 and a multitude of avx, and those take longer to run on an 8350 than the sse3's:-/ How the server got the idea my rig could handle avx in quantity is beyond me, even before the crash I never had so many WU sent to me.. Hell there are 28 avx at an estimated 3h17, and they in fact take 3h42 to complete:-( Regards, Cliff |
Send message Joined: 1 Jan 14 Posts: 300 Credit: 32,264,344 RAC: 7,374 |
Lower your cache size first or you will just get them all back again! Yes, it depends on the Server side software version of Boinc the project is running and some settings, several projects do it thinking the units were aborted accidentally and you really do want to crunch them. Sending you your old units back to you is no different than sending you new units, it's just programming, especially since you are asking for more units anyway. |
Send message Joined: 1 Jan 14 Posts: 300 Credit: 32,264,344 RAC: 7,374 |
Hi Mikey, I went into my own settings and don't see how to select which units you want to crunch, so it may just be 'assuming' that since you have the capability you want to crunch them. If you don't want to crunch gpu units, for instance, you may have to use an <exclude_gpu> line in a cc_config.xml file. |
Send message Joined: 19 Nov 14 Posts: 93 Credit: 30,066,240 RAC: 0 |
Hi Mikey, I only do a@h cpu tasks, my problem is that the a@h server keeps sending me avx tasks which take 3h23m to complete, sse3 tasks take 2h30m each.. I reset the project in the hopes it would send some sse3's but I still got avx, but at least now its not 20+, just 2:-) Regards, Cliff |
Send message Joined: 1 Jan 14 Posts: 300 Credit: 32,264,344 RAC: 7,374 |
Hi Mikey, You could probably do some config file stuff to only get the sse3 units if you really wanted too. |
Message boards :
Problems and bug reports :
Task Overload