21) Message boards : Number crunching : Fairer distribuiton of work(Flame Fest 2007) (Message 16439)
Posted 28 Feb 2007 by larry1186
Post:
OK, I am a newbie at LHC. Yet, I must admit to being thoroughly perplexed by this obsession with “fair” distribution. I just let BOINC poll for work and forget it. If work is assigned, wonderful. I have no plans to open a vein if there is no work available today, tomorrow, or even for the next several months. There are other projects, and mountains of numbers in need of crunching.

Those clamoring for “fairness” appear more interested in self-gratification than science. My grandchildren regularly ask for change to purchase an ice cream at the corner store. Sometimes I have some. Most times I do not. While they might express mild disappointment, they do not demand I alter my routine so there will always be coins available. They understand the true meaning of fairness

However, I do find the suggestion for placing limits on distribution of WUs intriguing. Perhaps I can convince the local casino to implement such a scheme. No other gamer at the roulette table, having once won, could be paid until my number came up on the wheel.

Please don't take any offense to my comments since none was intended. Good day, and welcome to the project!
[Disclaimer]I had fun thinking about this analogy, ask my wife, I like analogies and sometimes they get pretty far out there[/Disclaimer]
I think you may be a bit misguided on what is happening. To relate to your grandchildren/ice cream scenario: Let's say you have 7 grandkids, and you come across $5 and each ice cream costs $1. How do you distribute it?

What is happening: The first child that asks you for money gets it all. But only spends $1 today, $1 next week, and so on... You have an agreement with your source of $5 (I won't ask what that is... :) ) that you will not get any money to distribute to your grandkids until the previous distribution is spent. That one child will happily get an ice cream every week for 5 weeks and the other six kids get none.
The next $5 comes around and the first child to ask you for money (randomly different from the previous distribution) gets it all... and so on. I would like you to convince me (or any child) that this scenario is "fair"

What would be "fair": The first child that asks you for money gets $1 and the next kid gets $1, etc. but now only 5 kids get ice cream this week, the other two don't (but being good kids, they are ok with that and understand grandparent aren't made of money). Next week you get another $5 because the previous $5 was spent. Once again the kids ask you and the first 5 (again randomly different from the 5 last week) to ask get $1.

Then you have the crazy kids that "double fist" the ice cream cones (dual cores) that need two at a time, or the really obese ones that eat an ice cream cone every day (super fast over-clocked water-cooled), or the humble ones that come from a poor family and take a bite each day and put it back in the freezer til tomorrow and won't finish it until like 3 weeks later (the PII 233 MHz hosts). And no, this is not directed toward you grandkids at all. :)

When there isn't enough for everyone, the best way is to hold back on giving it on a first come first serve basis. Give a couple WUs to the first computers that ask and if they finish those, they can ask for more. This will benefit the project because you won't have a single host sitting on hundreds of WUs that won't finish within a month while my host has none and could be crunching those that are sitting idle on the "greedy" host. I realize that the users are not usually greedy, it's just the nature of the algorithm that a host will download hundreds of WUs since it's been starved for the last month or two or more. Right now there are ~3000 WUs out there somewhere and my host has not had a WU for some time now. It could have completed several WUs since the batch was distributed. That is what I don't find to be "fair", it is NOT fair to the project. I am not bothered by the fact that I don't have any LHC WUs to crunch. I am bothered by the fact that there are WUs to crunch and I feel helpless because I can't crunch them. But, such is life. @ Da Phoole: I agree with you, if there's work on my comp, fabulous; if not, oh well maybe next time.
22) Message boards : Number crunching : Fairer distribuiton of work(Flame Fest 2007) (Message 16343)
Posted 14 Feb 2007 by larry1186
Post:
I leave BOINC up, with allow new tasks enabled for LHC. Good, that's how you get WUs.

I looked in messages tab, my PC is periodically looking for work from LHC but no work found. Yup, that's how it works around here.

LHC has been setting "deferring communications" from from 7 sec to 2 hours. That's normal BOINC behavior.

I suspect those WU were available in only one of those 2 hr slots, and I missed them. Sometimes the work lasts a mere seconds or minutes, so getting in on the batch of WU takes a certain level of luck.

LHC is set to defaults - the only button on the BOINC Projects screen for it links to LHCathome -probably since I haven't run any WU's yet for it. Nope, that's the only link the project provides, ever.
23) Message boards : Number crunching : Please note: this project rarely has work (Message 16342)
Posted 14 Feb 2007 by larry1186
Post:


Bumped for your viewing pleasure.

To new admins (BTW, you're doing great so far): As River~~ mentioned earlier, could we get this thread stickied? (eeeewww) Maybe even a permanent disclaimer on the front page about the frequency of work?
24) Message boards : Number crunching : LHC credits not showing up in cross-project stats (Message 16341)
Posted 14 Feb 2007 by larry1186
Post:
Where is my pony? :*(


Steve Cressman has your pony! mwuahahaha... oh, sorry.
25) Message boards : Number crunching : LHC credits not showing up in cross-project stats (Message 16323)
Posted 13 Feb 2007 by larry1186
Post:
Hear "he-er" here!
26) Message boards : Number crunching : Could I ask for an update on the frontpage? (Message 16209)
Posted 31 Jan 2007 by larry1186
Post:
and is drunk in tiny glasses, preferably accompanied by a date


Whenever I have a date I tend to get drunk, but I don't wear glasses, at least not tiny ones.... Ni!
27) Message boards : Number crunching : Could I ask for an update on the frontpage? (Message 16195)
Posted 25 Jan 2007 by larry1186
Post:
It's not a problem guys (and gals, I assume), I know it's frustrating to be not kept up to date on things but believe me we are working on this as hard as we can..... oh look time for a tea break (I'm thinking of moving to work for the council ;-P)


All you have to do is set up a webcam in the room you're working in so we can keep an eye on you! And have another one on the screen so we can see what you're actually doing :) Then you don't have to worry about typing up something for updates, cuz that takes away your tea time ;)

But seriously, thanks for taking this project under your wing. It can't be easy pleasing 100,000+ (how many users are there anyway?) users
28) Message boards : Number crunching : Welcome QMUL Admins (Message 16152)
Posted 18 Jan 2007 by larry1186
Post:
The new admins better watch out for The Knights Who Say Ni! because:


Because... broken image?


Because, we will break your image if you do not appease us!
29) Message boards : Number crunching : Welcome QMUL Admins (Message 16120)
Posted 12 Jan 2007 by larry1186
Post:
Would anyone object to "No txt spk" being added to the box of rules on the left of page when you're posting a message?

u r the admin, u put da rulz ere. Personally I would like that rule, 'txt spk' sucks :)


I must agree. My opinion:I find it childish and annoying and it has no place in a scientific project forum.

[edit]I would like to add that shorthand (e.g. BTW) if perfectly ok, IMHO[/edit]
30) Message boards : Number crunching : Past Due Date (Message 16105)
Posted 10 Jan 2007 by larry1186
Post:
(I remember reading something about this, my apologies for not remembering where)

... like earlier in the thread... :)


Just making sure you were paying attention :) (to self: wow do I feel stupid) Thanks for jogging my memory!
31) Message boards : Number crunching : SOME greedy users (Message 16101)
Posted 10 Jan 2007 by larry1186
Post:
Now 6*4=24*60=1440 sec.


(6 WUs)*(4 hrs/WU)*(60 min/hr)*(60 sec/min) = 86400 seconds of work
32) Message boards : Cafe LHC : Greetings! (Message 16099)
Posted 10 Jan 2007 by larry1186
Post:
This is an awesomely lonely place!


No kidding! How ever many billion stars and solar systems out there and we haven't even exchanged "hello" with anyone...
33) Message boards : Number crunching : SOME greedy users (Message 16098)
Posted 10 Jan 2007 by larry1186
Post:
I don't know if there is a posibility for the project to relaunch WUs when the number of remaining results decay below a certain limit, say 20% or 10%. In this way, crunchers that are waiting for work will get WUs and they will finish the job in hours instead of days. All happy, the project finish sooner and crunchers have more work. :)

It is possible, increasing the replication for all workunits that haven't reached quorum yet.


SIMAP has done this. Initial replication = 2, quarum = 2, a couple days before the deadline they reissue the WUs that haven't reached quorum yet. From a servers point of view, it isn't doing much since there aren't any more results to send out, and the rush of incoming results is over so there is just one more surge (small, kind of like when the first batch of WUs reach deadline) before the deadline is reached.

[edit]added link, reworded[/edit]
34) Message boards : Number crunching : Past Due Date (Message 16097)
Posted 10 Jan 2007 by larry1186
Post:
A useful option would be for the core client to ask the server if a quorum exists before starting a new result. That way, an unneeded result could be aborted before computation started and a fresh replacement downloaded (if available). However, I imagine this would put a lot of extra strain on the servers and therefore might not be feasible. It would be very useful however if a machine has had to be switched off for a day or two and the cached work units were therefore a bit stale. Aborting stale unneeded results and moving on to something more useful would seem to be a good idea.


(I remember reading something about this, my apologies for not remembering where) I believe a feature such as this is in the works, whereas the client side support is included in 5.8.x but the server side needs worked on(?). So it isn't implemented yet. Only those WUs on a host that are in "ready to run" state will be eligible for the quorum-met-so-abort-remaining-WUs action. If a WU even has just 1 second of CPU time it will NOT be aborted by BOINC since that would involuntarily waste a users CPU time. The only thing wasted is bandwidth of downloading the unneeded WUs and some CPU time for communicating.

In my opinion, if the project needs results quickly, they should implement a small "in progress" cache per CPU (so people that download a lot of WUs don't hold on to them while others who are finished with theirs have empty caches) and use the abort-if-quorum-met feature (when available). The servers should be pushed to what they can handle and that is where the bottle-neck will occur instead of on the client side.
35) Message boards : Number crunching : SOME greedy users (Message 16050)
Posted 5 Jan 2007 by larry1186
Post:
Yank, another possible solution to consider: Since you're only crunching the two projects (SETI and LHC), could you set a large resource share in favor of LHC (1000:1 or 100:1) effectively using SETI as a backup project. Then when LHC does have work that's all that host will crunch and it will only take as long as your cache is set for to finish them (since it will switch to LHC full time because of high resource share).

I crunch 5 other projects aside from the two intermittent projects (LHC and SIMAP). LHC and SIMAP have 1000 resource share each and the other 5 projects have 100. So when neither LCH or SIMAP have work, each of the 5 still get equal shares. When either LHC or SIMAP have work, it crunches those full time. If both have work then they each get 50% (roughly).
36) Message boards : Number crunching : New Flood of WU's (Message 16034)
Posted 4 Jan 2007 by larry1186
Post:
Does this means that there were 17661 individual work units or was it 17661/5 or *5 units in all as each work unit is sent out 5 times?


There were 3532 (~17661/5) different WUs to crunch at that moment. Each of the 3532 WUs were copied 5 times and sent out. (I just noticed the terminology on the main page is a little misleading stating as "xxxxx workunits in progress" while the same number is reported on the server status page as "xxxxx results in progress")
37) Message boards : Number crunching : Not enough Disk space? (Message 15999)
Posted 3 Jan 2007 by larry1186
Post:
Some points to consider that commonly get overlooked:

Have you recently changed your settings on any of the projects (the most recently updated prefs get used) or on your account manager (if you use one)?

I would suggest double-checking where your prefs are being loaded from by looking in the "Messages" pane.

(I had a similar issue with WCG and freaked out when it asked for 17.5 GB only to find the prefs were not being loaded from where I thought they were)
38) Message boards : Number crunching : when we have news woks? (Message 15142)
Posted 19 Oct 2006 by larry1186
Post:

I dont know whats up but i have been trying to get WU now for more then one week and there has been zilch up till 30 seconds ago. Would you please solve this problem or let us know that we should shift our butts to some other project


"No work" is not a problem on this project, see this thread

Consider yourself lucky if you do get a WU (or 43). Happy crunching.


Previous 20


©2024 CERN