Optimal distribution of work

peristergiou
peristergiou
Joined: 22 Mar 05
Posts: 1
Credit: 351,669
RAC: 0

> Various solutions were

Message 9204 in response to message 9199

> Various solutions were posed, Like...
> 1. generating numbers to see who usually returns WUs and who goes over (can
> be caused by resetting or detaching)
> 2. or how fast WUs are usually returned (Average turnaround time...that was
> mine). Once you know this, if you have to re-issue a WU to a 4th machine,
> choose a fast turnaround machine.
> 3. Issuing to 4 hosts (as often happened anyway), and when 3 return it grant
> credit then.
>
> They chose #3.
>

So the policy of sending the WU's to four people but giving credits based on the first three to return them is simply a means of reducing time until pending credit is awarded? I am just asking that because the credit the fourth person gets has in the end nothing to do with how much time he/she spent on the WU. Since the middle of the three values is kept as the awarded credit, it is obvious that waiting for the fourth person to complete would most likely affect the awarded credit, so why not use that as well? Is there any other reason apart from awarding credit faster? I am guessing that over a period of time the credit I "win/lose" by being the fourth person will be balanced by the credit I "win/lose" by being one of the first three, but I was just curious...

Andreas.

Darren
Darren
Joined: 18 Jan 05
Posts: 94
Credit: 69,632
RAC: 0

> So the policy of sending

Message 9205 in response to message 9204

> So the policy of sending the WU's to four people but giving credits based on
> the first three to return them is simply a means of reducing time until
> pending credit is awarded? I am just asking that because the credit the fourth
> person gets has in the end nothing to do with how much time he/she spent on
> the WU. Since the middle of the three values is kept as the awarded credit, it
> is obvious that waiting for the fourth person to complete would most likely
> affect the awarded credit, so why not use that as well? Is there any other
> reason apart from awarding credit faster?

Actually, with the way einstein forms the quorum, the fourth result is only a "spare", just in case someone doesn't return one. Einstein only requires 3, but in their decisions for what circumstances they wanted to be prepared for, apparently they felt it was a good probability 1 of the crunchers might not complete the work - so they initially send it to 4 and then if 1 doesn't come back it still meets quorum just with the other 3.

Anything coming in after the quorum is met simply gets the "already determined" points awarded to it, and is not used in making any changes to the points that will be granted.

Keck_Komputers
Keck_Komputers
Joined: 18 Jan 05
Posts: 376
Credit: 5,744,955
RAC: 0

> > How long will it take

Message 9206 in response to message 9199

> > How long will it take them to rerun every tape from the old client
> (classic v2.xx) and all the tapes with the astropulse program?
>
> Not certain they will re-run every tape...could you post a link to where
> project heads stated that goal?
>
Sorry just saw it in discussions somewhere way back. And it was not every tape just every tape before the classic version that did better science came out.

> In any case, yes, once work is used up, seti will not re-run it.
> They have 3 projects in the wings to produce even more data though, and these
> were the primary reasons they created BOINC.
>
> Project 1 - Astropulse: Relook at each WU with an entirely different crunching
> program to look for evaporation of possible black holes. I haven't ever run
> this, but I suspect the time for each WU would be much less.
>
Actually the time per workunit is about double the regular seti workunits. However they get much fewer workunits per tape. As low as 100 to 1.

> Project 2 - Multi-beam receiver at Arecibo: A new antenna is/has be mounted
> on the Arecibo gimbol arm, along with the original single beam. Along with
> this will go a new faster/better recorder to make the tapes. This signal will
> have much more information than current tapes/WUs, and will use a wholly
> different crunching application to process.
>
> Project 3 - Extending seti detection to other dishes. Either getting data that
> is already collected by other dishes, or adding antennas to other dishes.
> (Example: Right now Arecibo does southern hemisphere and Northern isn't
> covered)
> -----
>
You have that backwards Arecibo covers part of the northern hemisphere and the southern is completely ignored. The Parkes obsevatory in Austrailia will be able to cover the southern hemisphere if & when.

> A little WU result issuing history:
> Originally in BOINC seti issued 3 WUs...there was much complaining in the
> forums that "2 people have returned the result, guy #3's machine has a history
> of not returning WUs (going over)...so my machine won't get credited until the
> WU is re-issued and a 4th guy sends it back (unless he too goes over)".
>
> Various solutions were posed, Like...
> 1. generating numbers to see who usually returns WUs and who goes over (can
> be caused by resetting or detaching)
> 2. or how fast WUs are usually returned (Average turnaround time...that was
> mine). Once you know this, if you have to re-issue a WU to a 4th machine,
> choose a fast turnaround machine.
> 3. Issuing to 4 hosts (as often happened anyway), and when 3 return it grant
> credit then.
>
> They chose #3.
>
Both #1 & #2 are being worked on though. It was alot easier to change #3 since quorum is a parameter of the work generating routine and could be changed daily if desired.

BOINC WIKI

BOINCing since 2002/12/8

Astro
Astro
Joined: 18 Jan 05
Posts: 257
Credit: 1,000,560
RAC: 0

I don't think the decision to

I don't think the decision to go from 2 of 3 to 3 of 4 was just to get users their credit faster. It also clears the WU and its associated data from the server DB faster, thus increasing the DB's ability to do more work.

lysdexia
lysdexia
Joined: 9 Mar 05
Posts: 97
Credit: 17,013
RAC: 0

cashes -> caches out paced ->

cashes -> caches
out paced -> outpaced
gimbol -> gimbal
it's -> its
seperate -> separate
tweeks -> tweaks


"My other computer is a virus farm."

wijata.com
wijata.com
Joined: 11 Feb 05
Posts: 113
Credit: 25,495,895
RAC: 0

> cashes -> caches > out

Message 9209 in response to message 9208

> cashes -> caches
> out paced -> outpaced
> gimbol -> gimbal
> it's -> its
> seperate -> separate
> tweeks -> tweaks

Oh c'on. Didn't You get the point of
this thread?
It's really useless here. Can't You use Your time better way. maybe do some open source programming?
And Your arrows (word -> anotherword) does nothing without context.

Try fix this: wiêkszoœæ ludzi tutaj jest znu¿ona Twoimi poprawkami, które i tak nic tu nie znacz¹...[PL]

ben
ben
Joined: 9 Feb 05
Posts: 36
Credit: 1,663
RAC: 0

> Both #1 & #2 are being

Message 9210 in response to message 9206

> Both #1 & #2 are being worked on though. It was alot easier to change #3
> since quorum is a parameter of the work generating routine and could be
> changed daily if desired.

David already has my code fixes for some of these, but credit etc is low priority to other aspects of boinc. These are scheduled for phase 3 after classic is closed everything settles down from that.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.