Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 12
Posts: 12   Pages: 2   [ Previous Page | 1 2 ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 3462 times and has 11 replies Next Thread
WaltSumner
Cruncher
Joined: Mar 14, 2008
Post Count: 5
Status: Offline
Reply to this Post  Reply with Quote 
Re: Should I be worried?

The slow machines are a G4 MacBook and a G5 iMac, both running OSX, 10.4.

One of the slow tasks finally validated. The wingman took 20 hours and finished on 12/12; my copy was done on the G4, which took just under 36 hours, and also finished on 12/12. Both were sent on 12/9. So slow machines do at least sometimes get matched and finish at roughly the same time. Probably Sekerob is right, they always get matched eventually, it is just hard for me to detect. There remain one task each that my G4 (sent 12/7) and G5 (sent 12/6) Macs finished on 12/9 that are still "waiting to be sent" to a wingman, and one G4 task sent 12/4 and completed 12/7 where the wingman is "in progress", so the matched queues apparently are not strictly FIFO by send or return date for these slow machines. Interesting...

My faster Intel Macs are getting matched to even faster machines, and both reply within 24 hours of one another. The only exception has been one "no response": the task was immediately sent to another fast machine that responded within a few hours.
----------------------------------------
--WS
[Dec 14, 2009 4:52:44 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: Should I be worried?

The techs have a dilemma. If insisting on matching through the last WU, if that what it is, the batches would complete with substantial delay... a HCMD batch e.g. needs 30 days I read.

There is no official word at all if there was a broader 'CPU matching' implementation i.e. it's maybe something still being trialed and tuned in the production stream. We know from past that the Homogeneity group on Mac for HCC was just Intel and PPC, so there is no absolute requirement... and still a fast one that runs with zero buffer doing 24/7 could get a wingman device that has a 3 day buffer and then goes on a week long field trip. There will be oddities on an active member count of many thousands, and staring at BOINC will bring out it's worst character ... it's alive ;>)
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
----------------------------------------
[Edit 1 times, last edit by Sekerob at Dec 14, 2009 8:01:51 AM]
[Dec 14, 2009 8:01:24 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 12   Pages: 2   [ Previous Page | 1 2 ]
[ Jump to Last Post ]
Post new Thread