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: 388
|
![]() |
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Do you think that you'll get 1 day work of HCMD2, NOW? If you have 4 days work still, what's the complaint?
It's only a suggestion based on the fairly constant [with bounds] runtimes I'm observing for most sciences at WCG and proposing how to scrape a few in the end run which will not do any big inflating would they run longer and certainly at 1 day cache will not push the envelope to your 12+ days ;>). You know your machine, you set the prefs as the way your machines deals with variability! --//-- |
||
|
astrolabe.
Senior Cruncher Joined: May 9, 2011 Post Count: 496 Status: Offline |
All Done: Help Cure Muscular Dystrophy - Phase 2 7,703,044 24,375 10:004:19:59:55
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Done in Maggiore
![]() --//-- |
||
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Do you think that you'll get 1 day work of HCMD2, NOW? If you have 4 days work still, what's the complaint? It's unlikely you'll get much if any HCMD2-work now, but I was trying to point-out that for a computer with little/no HCMD2-work, your method of caching "1 day" of other WCG-work likely would be too little actual work to not be idle before can re-fill with more WCG-work in 24 hours, since there is little/no HCMD2-work to get. ![]() "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
OK, OK... I get it, I get it ! ...to avoid the HCMD2 idle times. It was early and I hadn't had my first cup of coffee.
----------------------------------------CJSL |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Most will be hands on enough to be able to tune the value without ending up with loads of ''other''. DCF goes down slow, so any incidental quickies wont shrink the cache dramatically and any long wont inflate much as the DCF will in main be driven by the 'other'. Plz open new thread if you wish to extend the debate on what's best, so it can be put to good use by others beyond this HCMD2 crunching thread... asked you before... draft an FAQ with best practices from your vast experience.
------------------------------------------//-- [Edit 1 times, last edit by Former Member at Feb 17, 2012 2:52:21 PM] |
||
|
coolstream
Senior Cruncher SCOTLAND Joined: Nov 8, 2005 Post Count: 475 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Sekerob, I respect what you are saying about a new thread for this topic being a good idea. One 'additional tool' that makes micromanagement much easier as you know is BoincTasks. I think that for people who are trying to squeeze the last few drops out a project, being able to see the bigger picture really helps when attempting to make tweaks.
----------------------------------------I went to BoincManager on one machine recently with a view to doing a final tweak and it felt like I was in a submarine that had no periscope. Going back to BoincTasks, I was much more able to see a trend in the distribution of those elusive final WUs. I hope that when you are writing your FAQ, you will make some mention of BoincTasks. ![]() Crunching in memory of my Mum PEGGY, cousin ROPPA and Aunt AUDREY. |
||
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3715 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I can confirm that SekeRob's scenario works well: four days before his post I have used the same one while I was finishing a stack of grand children WUs (i.e. low DCF at this time) and, suddenly, within a few minutes, my quad has received scores of repair** work, half children and half parents (some back to 21xx batches!), the parents with their outrageously underestimated durations as usual.
----------------------------------------Finally, with a lot of micromanagement I could do the work in 4 calendar days as requested, the last three parents completing 2-3 hours after the deadline but much before the replacement jobs had the smallest chance to come back. ![]() And the good news for batch hunters is that several parents were so tough that they have been stopped at 6 hours (usually this quad completes all parents in 5-8 hours), so there are probably still many chidren and grand children to chase in the coming days. ** Considering the batch numbers I have seen maybe it was more "housekeeping-by-the-techs" than true repair works, but they were all 4-day jobs anyway. ![]() ---------------------------------------- [Edit 1 times, last edit by JmBoullier at Feb 18, 2012 1:17:15 PM] |
||
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1677 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
For my-self, I experience currently a crazy situation.
----------------------------------------Since I was business traveling last week, I decided to increase my cache from 1.5 days to 5 days. I did it carefully step by step: +.5 day/12 hours. Unfortunately, the last project update occurred by two hosts just after a large collection of very small WUs (<1 hour/WU). The consequence is that these both hosts fetched a lot of WUs at once which seem to take much more time (3 to 5 hours) than Boinc did expect. It looks like, that I will probably send back many WUs over the deadline. Within 5 years, it is the first time I experience such a mistake. Additionally, boinc is reporting a cache size varying from hour to hour between 7 days and 25 days, without downloading new WUs since 4 days. Something went wrong. Cheers, Yves |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
One day left!
|
||
|
|
![]() |