Message boards : Number crunching : Fairer distribuiton of work(Flame Fest 2007)
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 10 · Next

AuthorMessage
Daxa

Send message
Joined: 29 Dec 06
Posts: 100
Credit: 184,937
RAC: 0
Message 16268 - Posted: 7 Feb 2007, 2:47:40 UTC
Last modified: 7 Feb 2007, 3:23:49 UTC

This is a bit of an anachronism, but in terms of "fairer distribution of work", one should seriously consider participating in a BOINC project called XtremLab.

It actually studies WU allotment and overall grid efficiency. It has the potential to fix a LOT of the work distribution problems in most, if not all, BOINC-based projects.

Check out this link for more info. Account Creation is not as straight-forward as with other projects, but it's not difficult to get started. I truly believe this is a valuable project for the future of Distributed Computing.



_______

"Three quarks for Muster Mark!"
. . . . . . . - James Joyce, Finnegans Wake . . . .

ID: 16268 · Report as offensive     Reply Quote
Profile Bober [B@P]
Avatar

Send message
Joined: 10 Apr 06
Posts: 7
Credit: 69,019
RAC: 0
Message 16284 - Posted: 9 Feb 2007, 23:17:00 UTC - in response to Message 16268.  

This is a bit of an anachronism, but in terms of "fairer distribution of work", one should seriously consider participating in a BOINC project called XtremLab.

It actually studies WU allotment and overall grid efficiency. It has the potential to fix a LOT of the work distribution problems in most, if not all, BOINC-based projects.

Check out this link for more info. Account Creation is not as straight-forward as with other projects, but it's not difficult to get started. I truly believe this is a valuable project for the future of Distributed Computing.



And one should think twice before crunching for XtremLab, as admins there have not given a word about their work since months. We (participants) have no idea about project state, results...complete silence :(
ID: 16284 · Report as offensive     Reply Quote
Profile clownius
Avatar

Send message
Joined: 1 May 06
Posts: 34
Credit: 64,492
RAC: 0
Message 16285 - Posted: 10 Feb 2007, 3:53:10 UTC

The easiest fix i could think of here is something RCN was doing for a while when they couldnt generate Wu's as fast as we could crunch them. There was a limit on the number of WU's a computer could have at any time. Once you returned one you could have another. Seems a fair method to me as faster crunchers get more work and the work is returned quicker this way.
As for Xtremlab is crunched it for a while and about the only think i achieved was some credits. Each WU is just 10mins long and often shorter and achieves little other than keeping your CPU warm as far as i can tell. The admins appear silent or MIA and it seems to me only good for those using those modified BOINC clients to claim 100's of credits per hour without anyone really noticing.
On the first point i may have posted that in the past but i cant remember and its something the new admins may want to look into once the project moves.
ID: 16285 · Report as offensive     Reply Quote
Profile WimTea

Send message
Joined: 27 Mar 06
Posts: 21
Credit: 1,731
RAC: 0
Message 16286 - Posted: 10 Feb 2007, 8:34:08 UTC
Last modified: 10 Feb 2007, 8:34:45 UTC


The easiest fix i could think of here is something RCN was doing for a while when they couldnt generate Wu's as fast as we could crunch them. There was a limit on the number of WU's a computer could have at any time. Once you returned one you could have another. Seems a fair method to me as faster crunchers get more work and the work is returned quicker this way.


A similar sugggestion was already discussed in this same thread at http://lhcathome.cern.ch/forum_thread.php?id=2286&nowrap=true#15243. Anyway, it's nice to see other people coming up with the same general idea.
ID: 16286 · Report as offensive     Reply Quote
Daxa

Send message
Joined: 29 Dec 06
Posts: 100
Credit: 184,937
RAC: 0
Message 16289 - Posted: 10 Feb 2007, 23:14:48 UTC - in response to Message 16284.  

This is a bit of an anachronism, but in terms of "fairer distribution of work", one should seriously consider participating in a BOINC project called XtremLab.

It actually studies WU allotment and overall grid efficiency. It has the potential to fix a LOT of the work distribution problems in most, if not all, BOINC-based projects.

And one should think twice before crunching for XtremLab, as admins there have not given a word about their work since months. We (participants) have no idea about project state, results...complete silence :(

This is an excellent point, and a fact I was not aware of. I'm especially unnerved by this because there is some doubt that XtremLab is COMPLETELY not-for-profit. Anyone who has information about the profit status of XtremLab please post... I'm consider detaching from the project (and yes, I am the one who was just recommending it! Oh well...)

_______

"Three quarks for Muster Mark!"
. . . . . . . - James Joyce, Finnegans Wake . . . .

ID: 16289 · Report as offensive     Reply Quote
Profile Kibble

Send message
Joined: 14 Jan 07
Posts: 33
Credit: 255,657
RAC: 0
Message 16291 - Posted: 11 Feb 2007, 17:46:11 UTC

As I'm a new player here, without credit yet, my take on the issue may be a bit different from the majority of the posters. Personally I'm inclined to recommend waiting until the project is up and running before addressing the issue of fairness in the downloading of work units. When the boffins want to do big science and the infrastructure is available, I'm sure there will be many projects which means many many flops will be required. The LHC is gonna be too expensive to just run a few experiments then shut down. More people will be available to make changes in the servers' progaming when things start moving if needed.

I just signed up a short while ago simply due to being interested in watching the scientific discoveries. If the few cycles I have available can make things happen any quicker then I'll be quite happy.

There are quite a few other projects available for the result junkies, and I also run both SETI and Einstein programs currently.

ID: 16291 · Report as offensive     Reply Quote
Ravens

Send message
Joined: 14 Dec 06
Posts: 29
Credit: 128,225
RAC: 0
Message 16292 - Posted: 11 Feb 2007, 21:04:47 UTC - in response to Message 16291.  

As I'm a new player here, without credit yet, my take on the issue may be a bit different from the majority of the posters. Personally I'm inclined to recommend waiting until the project is up and running before addressing the issue of fairness in the downloading of work units. When the boffins want to do big science and the infrastructure is available, I'm sure there will be many projects which means many many flops will be required. The LHC is gonna be too expensive to just run a few experiments then shut down. More people will be available to make changes in the servers' progaming when things start moving if needed.

I just signed up a short while ago simply due to being interested in watching the scientific discoveries. If the few cycles I have available can make things happen any quicker then I'll be quite happy.

There are quite a few other projects available for the result junkies, and I also run both SETI and Einstein programs currently.


I too am new - joined before Christmas and have never had a WU yet, so I haven't even seen the graphics. Would be nice if newcomers got one WU on joining, like a housewarming present. I leave BOINC up most of the time, maybe next time LHC has work I'll get some before the feeding frenzy starts.
In the meantime, I too have other BOINC projects chugging along.
ID: 16292 · Report as offensive     Reply Quote
Bob Guy

Send message
Joined: 28 Sep 05
Posts: 21
Credit: 11,715
RAC: 0
Message 16309 - Posted: 13 Feb 2007, 3:51:19 UTC

About a half hour ago I managed to get 25 WUs but there were no more available by the time I got that many downloaded. Unfortunately I'm finishing them in about 6 minutes each.
ID: 16309 · Report as offensive     Reply Quote
Ravens

Send message
Joined: 14 Dec 06
Posts: 29
Credit: 128,225
RAC: 0
Message 16326 - Posted: 13 Feb 2007, 18:26:01 UTC - in response to Message 16309.  

About a half hour ago I managed to get 25 WUs but there were no more available by the time I got that many downloaded. Unfortunately I'm finishing them in about 6 minutes each.

I'd like to see just 1 WU, to get started. I leave BOINC up, with allow new tasks enabled for LHC. I looked in messages tab, my PC is periodically looking for work from LHC but no work found. LHC has been setting "deferring communications" from from 7 sec to 2 hours. I suspect those WU were available in only one of those 2 hr slots, and I missed them.
How do I set BOINC to try LHC more often? Not every few seconds, but maybe every 15 minutes? I have Einstein and Rosetta and Seti in here too - all at 100% share and CPU at 30% usage (I have a lap top).
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. I have BOINC 5.8.11 on WinXP.
Thanks in advance - Mike
ID: 16326 · Report as offensive     Reply Quote
PovAddict
Avatar

Send message
Joined: 14 Jul 05
Posts: 275
Credit: 49,291
RAC: 0
Message 16329 - Posted: 13 Feb 2007, 18:52:10 UTC - in response to Message 16326.  

How do I set BOINC to try LHC more often?

You don't. That would cause more server load. There is a reason why the connect interval goes higher.
ID: 16329 · Report as offensive     Reply Quote
Ravens

Send message
Joined: 14 Dec 06
Posts: 29
Credit: 128,225
RAC: 0
Message 16339 - Posted: 14 Feb 2007, 18:30:41 UTC - in response to Message 16329.  

How do I set BOINC to try LHC more often?

You don't. That would cause more server load. There is a reason why the connect interval goes higher.

Yes - I am aware of that.
I've noticed it's worse than I thought though. If I run any other application (Einstein or Rosetta for example) then the LHC aplication stops looking for work completely. Even if it had thought it was a 1 minute interval to the next "poll" - nothing - for hours. If I suspend the running applications LHC still doesn't look for work - I have to do an update on the project screen to get it back into it's periodic polling.
I realize this is likely a BOINC issue, but there are so many discussion groups in there - any suggestions as to where this would fit? Is it a problem or a feature?
ID: 16339 · Report as offensive     Reply Quote
KAMasud

Send message
Joined: 7 Oct 06
Posts: 114
Credit: 23,192
RAC: 0
Message 16340 - Posted: 14 Feb 2007, 19:24:56 UTC - in response to Message 16339.  
Last modified: 14 Feb 2007, 19:30:34 UTC



How do I set BOINC to try LHC more often?

You don't. That would cause more server load. There is a reason why the connect interval goes higher.

Yes - I am aware of that.
I've noticed it's worse than I thought though. If I run any other application (Einstein or Rosetta for example) then the LHC aplication stops looking for work completely. Even if it had thought it was a 1 minute interval to the next "poll" - nothing - for hours. If I suspend the running applications LHC still doesn't look for work - I have to do an update on the project screen to get it back into it's periodic polling.
I realize this is likely a BOINC issue, but there are so many discussion groups in there - any suggestions as to where this would fit? Is it a problem or a feature?


:-) dont worry as you can see from my RAC i am also floating in the same boat:-)
;-) this time got nothing?:-)
Regards
Masud.

ID: 16340 · Report as offensive     Reply Quote
larry1186

Send message
Joined: 4 Oct 06
Posts: 38
Credit: 24,908
RAC: 0
Message 16343 - Posted: 14 Feb 2007, 20:37:32 UTC - in response to Message 16326.  

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.
Don't get distracted by shiny objects.
ID: 16343 · Report as offensive     Reply Quote
Ravens

Send message
Joined: 14 Dec 06
Posts: 29
Credit: 128,225
RAC: 0
Message 16344 - Posted: 14 Feb 2007, 21:43:46 UTC - in response to Message 16343.  

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.


If you look at my post the next day (feb 14) on this - my problem is if I run any other applications, LHC doesn't do ANY checks.
ID: 16344 · Report as offensive     Reply Quote
Profile Steve Cressman
Avatar

Send message
Joined: 28 Sep 04
Posts: 47
Credit: 6,394
RAC: 0
Message 16346 - Posted: 15 Feb 2007, 0:12:45 UTC

I do several projects and my boinc manager continues to ask LHC for work. I get work from almost every batch. I just adjusted my boinc settings and the scheduler does the rest without my intervention. Sure there have been a couple batches that I missed out on because the work happened to come and go while I happened to be in a long back off. It goes through a cycle starting with minute backoffs up to a 4 hour backoff. Then the cycle repeats with ever increasing interval until it gets to about 4 hours again then it goes back to doing short backoffs again. Do you now see a pattern as to how it works? But you will have better success using a small cache, hint hint.

Steve
98SE XP2500+ @ 2.1 GHz Boinc v5.8.8
ID: 16346 · Report as offensive     Reply Quote
River~~

Send message
Joined: 13 Jul 05
Posts: 456
Credit: 75,142
RAC: 0
Message 16347 - Posted: 15 Feb 2007, 7:18:31 UTC - in response to Message 16339.  

If I run any other application (Einstein or Rosetta for example) then the LHC aplication stops looking for work completely. Even if it had thought it was a 1 minute interval to the next "poll" - nothing - for hours. If I suspend the running applications LHC still doesn't look for work - I have to do an update on the project screen to get it back into it's periodic polling.


There are four different things going on here, possibly in some combination.

The first is the normal polling backoff.

1. For a few times it tries at 1min intervals, but then slowly degrades in random steps (so it can sometimes get quicker for one or two tries) till it is trying at random intervals with a max of 4hrs. This process is unaffected by other work on the box. Please do not interfere with this process, when people do this in significant numbers it kills the server - that is why there is a backoff.

2. The second is the cache size & no work fetch logic. This varies from version to version of the BOINC client, but the client will not ask for work when it thinks your box already has enough. That is why I recommend a cache roughly equal to one work unit of your other projects - except CPDN ;-)

3. If LHC downloaded a lot of work last time it had some, and then ran in earliest deadline first to get it finished, it may be in negative debt mode. You can look for your debt in the client_state.xml file or with a sutiable debt reader program (remind me where to download it from please someone).

4. If the cleint could not contact LHC for five times running, it then backs off for either 24hrs or 1 week depending in the client version. This backoff is to prevent network load on a server returning from downtime. If you see this message in the message window, please check you can load the LHC front page and that it shows a plausible (possibly zero) value for work in progress - if you can't see the LHC front page or it complains of too many connections then please wait a while before trying the front page again. Please don't retry more than two or three times a day in this case or you become part of the problem. On the other hand if the long backoff is because of a known network problem local to yourself, then by all means retry after verifying that the LHC front page is OK. As you say, you can retry by clicking update.

How do you know which is which? Well in cases 2 and 4 there will be no effect in suspending all other projects, wheras in cases 1 and 3 suspending all other projects will trigger an update request within a few minutes (not always immediately).

You tell 2 and 4 apart by looking at the message window messages. You tell 3 by a message 'No Work Fetch' (remind me which window this is in please - in BoincView it is in the hosts window under network status, but I never use the official manager so can't remember where this message appears).

Hope that helps
R~~
ID: 16347 · Report as offensive     Reply Quote
Profile Steve Cressman
Avatar

Send message
Joined: 28 Sep 04
Posts: 47
Credit: 6,394
RAC: 0
Message 16354 - Posted: 15 Feb 2007, 13:41:33 UTC



BoincDebtViewer



Steve
98SE XP2500+ @ 2.1 GHz Boinc v5.8.8
ID: 16354 · Report as offensive     Reply Quote
PovAddict
Avatar

Send message
Joined: 14 Jul 05
Posts: 275
Credit: 49,291
RAC: 0
Message 16355 - Posted: 15 Feb 2007, 14:16:50 UTC - in response to Message 16347.  

2. The second is the cache size & no work fetch logic. This varies from version to version of the BOINC client, but the client will not ask for work when it thinks your box already has enough. That is why I recommend a cache roughly equal to one work unit of your other projects - except CPDN ;-)

On newer versions of BOINC, the setting actually started doing what the name says. It's NOT a cache size setting, it's a connection interval setting. And due to the way it works, sometimes having higher "Connect every X days" makes it download LESS work! Because of deadline calculations; "if you won't connect for 7 days, you won't be able to report all this in time, so I won't get more work". Big discussion about it here and here. Set it to how often you connect to the Internet. Always on connections should have it < 1.0.
ID: 16355 · Report as offensive     Reply Quote
Profile caspr
Avatar

Send message
Joined: 26 Apr 06
Posts: 89
Credit: 309,235
RAC: 0
Message 16399 - Posted: 24 Feb 2007, 5:36:31 UTC

I understand that people want to work the project,and I understand that they dont connect as often as they would like to be able to. But it seemes that as far as the project is concerned, they could set up some kind of quota system that would help them get their work finished sooner and give everyone a shot at at least a few wu's at least till they get finished with the migration and work becomes more avalible. ( I HOPE! )
A clear conscience is usually the sign of a bad memory


ID: 16399 · Report as offensive     Reply Quote
Profile John Hunt

Send message
Joined: 13 Jul 05
Posts: 133
Credit: 162,641
RAC: 0
Message 16400 - Posted: 24 Feb 2007, 7:21:56 UTC - in response to Message 16289.  

This is a bit of an anachronism, but in terms of "fairer distribution of work", one should seriously consider participating in a BOINC project called XtremLab.

It actually studies WU allotment and overall grid efficiency. It has the potential to fix a LOT of the work distribution problems in most, if not all, BOINC-based projects.

And one should think twice before crunching for XtremLab, as admins there have not given a word about their work since months. We (participants) have no idea about project state, results...complete silence :(

This is an excellent point, and a fact I was not aware of. I'm especially unnerved by this because there is some doubt that XtremLab is COMPLETELY not-for-profit. Anyone who has information about the profit status of XtremLab please post... I'm consider detaching from the project (and yes, I am the one who was just recommending it! Oh well...)


Sorry for delayed reply to this post!
XtremLabs admins are noted for their lack of communication and updates - last update on the front page was August last year.
On a separate issue, some crunchers have stopped doing this project because of the blatant 'overclaiming' of credits that was being done by some of the top crunchers. Admins were made aware of the situation, but nothing seems to have been done. On QMC, however, where a similar thing has been happening, the admins have stated publicly that something is going to be done - credits deducted! Nanohive is another project where this is happening.
Most of this overclaiming is due to the use of Boinc 5.5.0 which is infamous for overclaiming. On a project where quorum 2 is needed, all it takes is two results using 5.5.0 to get inflated credits. HashClash and Pirates have found a way to stop the use of 5.5.0; wish all projects could find a way to do this!


ID: 16400 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 10 · Next

Message boards : Number crunching : Fairer distribuiton of work(Flame Fest 2007)


©2024 CERN