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: 2370
|
![]() |
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Nice rain we got, I finally made my red badge! Decided to stop with that; I normally don't abort units when I reach my target, but in this case I made an exception and aborted about 30 unstarted WUs to make some fellow badge-hunters happy :)
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just noticed that since the middle of last night the sr sq sd sub-type results have started to arrive in queue with their own estimated run time , about 1/3rd of the 'p'. Before they showed the same time as the 'p'. Certainly makes cache total projection more accurate.
--//-- |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I've set my profiles for DDDT2 and 1/2 day cache, and I'm picking up a steady flow of work. Just 100 days to go for the blue, 18 cores working. Will this rain last long enough??? best guess.... my goal, 9 cores, +45 days, I WILL get mine, therefore you should also ;) |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
![]() ![]() ![]() ![]() |
||
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Nice rain we got, I finally made my red badge! Decided to stop with that; I normally don't abort units when I reach my target, but in this case I made an exception and aborted about 30 unstarted WUs to make some fellow badge-hunters happy :) You can afford to abort 30 units ![]() In general to get 30 units I need to wait for months. That is the difference between the well fed and the starving. ![]() ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ive noticed another issue here. My computer is running tasks and then will stop them, sometimes an hour or so in sometimes a few hours in, and put them in waiting to run,half way completed, and move onto a new task. Ive got a pile of half done tasks now. Is this normal?
----------------------------------------Im thinking of suspending the remaining ready to start tasks and making it go back to the half dones, that won't cause any problems will it? Aaron [Edit 1 times, last edit by Former Member at Jan 16, 2012 12:07:28 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Aaron,
----------------------------------------What's the deadline range [remaining time before due] of those tasks and what client version are you running? Also what are the "Connect Every xx days", Additional Buffer days and Switch between applications [default 60 minutes] set at? That latter is pumped by some to rush repair jobs ahead, but works into the total planning. Generally, if more work is cached than the client thinks is can complete, it will go into a seeming random mode of trying the latest received tasks and see how long they take. For instance, if your cache is 7 days, and there are tasks in there that are due in 6 days, you'd see such behavior... not specific to DDDT2. Many have queried this, and so happens I queried this today again with developers of the new test client. Got 5 days build up on one machine and no task is due until 8.5 days or later. The latest tasks are being processed before the ones that are due earlier [processing in High Priority as indicated in the Status column of BOINC Manager, task view]. "Normal" operation is FIFO, no matter what the deadline is. --//-- edit: Duration Correction Factor also goes into the equation. If a task is received with 4 hours estimated run time, but takes 6, than all other work in cache is assumed to also run 1.5 times longer. Suddenly a 5 day cache looks to BOINC as needing 7.5 days to complete. [Edit 2 times, last edit by Former Member at Jan 16, 2012 12:40:53 PM] |
||
|
astrolabe.
Senior Cruncher Joined: May 9, 2011 Post Count: 496 Status: Offline |
In general to get 30 units I need to wait for months. That is the difference between the well fed and the starving. Seems like you are choosing to starve. Your GFlops are 20x of mine and I have 290 dddt2 WU IP with a 1.3 day 'set and forget' cache. ![]() |
||
|
BSD
Senior Cruncher Joined: Apr 27, 2011 Post Count: 224 Status: Offline |
My dedicated Windows devices to this rain have been getting a steady inflow with a 2 day cache. I took the only dedicated Mac off this rain a few days ago due to it only getting 3 WUs, it is now cruching on other goodies. Surprisingly, my other Mac that I left crunching GFAM and DSFL with all intermittant projects (DDDT2 included) checked in the device profile, received 1 DDDT2 WU. What a nice surprise.
----------------------------------------![]() ![]() dg05_ a193_ sdb000_ 1-- 640 Pending Validation 1/15/12 14:00:24 1/16/12 02:33:15 1.07 22.9 / 0.0 dg05_ a193_ sdb000_ 0-- - In Progress 1/15/12 14:00:11 1/25/12 14:00:11 0.00 0.0 / 0.0 <-- Mine [Edit 1 times, last edit by BSD at Jan 16, 2012 5:49:10 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thank you sekerob.
----------------------------------------My main machine with 6 cores and a ton of ram is no problem. one of my others now which is doing this, started complaining about low virtual memory and having to increase the swapfile size so I let it. It just seemed odd that it would start a bunch of jobs run them a few hours then swap off. I can see it doing this once as you explained to work on 'rush jobs' but to do this again and again seemed odd. The virtual memory machine, I unplugged it from the internet for a while, suspended all the waiting to run jobs, and let it eat through the half done ones, then unsuspended them and plugged it back in to report. The reason i unplugged is that once before some time ago I suspended some jobs to let a few others (beta's ) take priority and boinc aborted every one immediately that I put in suspend. Given how long I waited to finally get a job for this project I'd have cried if that happened so took the ethernet cable out so it could not do that and everything would be put back to normal by the time I let it talk to the net again. again, thanks for the detailed reply. Edit: My earliest jobs are due on the 25th 7pm ish. and I tried to max the cache but it limited my jobs before i got there so I probably got 5 days worth saved up. Aaron [Edit 1 times, last edit by Former Member at Jan 16, 2012 10:30:44 PM] |
||
|
|
![]() |