Thoughts On Credits

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9352143
RAC: 0

RE: RE: Credit News

Message 83719 in response to message 83718

Quote:

Phew!
Quote:
Does this ring a bell for long time EAH Crunchers? ;-)

Yup. Been there, done that. The open source app question was resolved a while ago here at E@H. Also 'optimised' clients have been made irrelevant. I agree with Bikeman, let's leave the specifics of that one for the MW@H boards! :-)

Cheers, Mike.

Agreed, and that was the only point I wanted to make.

As far MW goes, they are a young Beta project which is pushing BOINC right to the edge of what is possible for the widely diverse, loosely coupled computing platform BOINC provides. Mistakes are going to be made.

As such, they deserve to be cut a wide margin of slack as they get up to speed with all aspects of computing in BOINCland.

Alinator

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6589
Credit: 318597304
RAC: 405039

RE: Agreed, and that was

Message 83720 in response to message 83719

Quote:

Agreed, and that was the only point I wanted to make.

As far MW goes, they are a young Beta project which is pushing the BOINC right to the edge of what is possible for the widely diverse, loosely coupled computing platform BOINC provides. Mistakes are going to be made.

As such, they deserve to be cut a wide margin of slack as they get up to speed with all aspects of computing in BOINCland.


I think what is/has been gradually happening to BOINC projects overall is a gradual convergence toward 'good' operational rules. This is both a social and a technical exercise. 'Good' is being progressively refined to represent what 'works' to further a project's goals in the long term as stable centres of human and technical productive interaction. I know that reads like some silly corporate mission statement but I think has the essence of lessons learnt.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9352143
RAC: 0

RE: ... I know that reads

Message 83721 in response to message 83720

Quote:

... I know that reads like some silly corporate mission statement but I think has the essence of lessons learnt.

Cheers, Mike.

LOL...

As much as I hate to admit that, even blind squirrels (corporations) find a nut occasionally! :-D

Alinator

Thunder
Thunder
Joined: 18 Jan 05
Posts: 138
Credit: 46754541
RAC: 0

RE: Understand, but to do

Message 83722 in response to message 83716

Quote:
Understand, but to do so, I'd have to attach...which would blow my semi-sorta "street cred" of not being someone that chased credits...since they would probably be the top credit awarder now that Cosmology has tanked...

You know, I have to agree that they're still paying too much, though they've had their own fair share of beta 'teething pains' in the last couple months (something that few take into account when they look at credit), but come on...

To *&%&^% with credits, the science is well... I want to use the expression "wicked cool"! ;) I know, I know... I should be more indignant about the credit thing, but the project is just soooo "up my alley" that I have to crunch it! It's honestly the first BOINC project at which I read the first paper released on it and actually understood about 90% of it. When I had questions about it, the lead scientist was giving me answers within hours of posting... and not just answers, but incredible explanations with pictures, etc.... I mean, how freakin' cool is THAT?! :D

Besides, you could always attach like 1 machine and give it a small resource share just to do a wee bit and be able to post, right? :)

(Okay, okay, I'll stop talking about some other project and getting totally off-topic) ;-)

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282700
RAC: 0

RE: As much as I hate to

Message 83723 in response to message 83721

Quote:

As much as I hate to admit that, even blind squirrels (corporations) find a nut occasionally! :-D

That's right, Silly rabbit, tricks are made for kids, don't you know that. You without me is like corn flakes without the milk! This is my world. You're just a squirrel trying to get a nut! Now get on outta here. Scat!

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282700
RAC: 0

RE: RE: Understand, but

Message 83724 in response to message 83722

Quote:
Quote:
Understand, but to do so, I'd have to attach...which would blow my semi-sorta "street cred" of not being someone that chased credits...since they would probably be the top credit awarder now that Cosmology has tanked...

You know, I have to agree that they're still paying too much, though they've had their own fair share of beta 'teething pains' in the last couple months (something that few take into account when they look at credit), but come on...

To *&%&^% with credits, the science is well... I want to use the expression "wicked cool"! ;)

I will assume that's about MW? Cosmology's level is down roughly "bottom-feeding" right now, along with LHC. I think Orbit currently is the bottom, but anyway...

The only problem I have with the "science" there being "wicked cool", is if it is so "wicked cool", then why does the admin apparently not check for valid results? I don't get the whole thing about not wanting redundancy. I just don't understand a "quorum of one"... :shrug:

Thunder
Thunder
Joined: 18 Jan 05
Posts: 138
Credit: 46754541
RAC: 0

RE: The only problem I have

Message 83725 in response to message 83724

Quote:
The only problem I have with the "science" there being "wicked cool", is if it is so "wicked cool", then why does the admin apparently not check for valid results? I don't get the whole thing about not wanting redundancy. I just don't understand a "quorum of one"... :shrug:

It has to do with the nature of their algorithm. They're using what they call Generic Maximum Likelihood Evaluation and it involves a "genetic" search, so the outcome of earlier tasks alter the parameters of future tasks generated (which is why they never have more than about 500 tasks ready to send). If they had to wait for a quorum it would actually throw things off worse. I can't explain it much more (that's among the 10% of the paper that I struggled with), but suffice to say that 'bad' results don't affect it much because the algorithm eliminates 'outliers' pretty well as long as there aren't too many of them.

You are correct in that it makes them one of the most vulnerable BOINC projects to being 'gamed' for credits though. :(

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6589
Credit: 318597304
RAC: 405039

RE: ... so the outcome of

Message 83726 in response to message 83725

Quote:
... so the outcome of earlier tasks alter the parameters of future tasks generated .....


Hmmmm .... what a dynamic that would/could generate. An actively/tuned search strategy programmatically applied, very much a different beast. Alas more spoof-able, and rather more sensitive to trends/happenstance. You'd have to admire them for trying that scheme out! :-)

That does not happen intra-run at E@H. Our strategy changes are made on a much longer time scale and with 'slow' humans explicitly in the loop. Committees, schedules, reviews, milestones etc. Very sedate by comparison.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9352143
RAC: 0

RE: Hmmmm .... what a

Message 83727 in response to message 83726

Quote:

Hmmmm .... what a dynamic that would/could generate. An actively/tuned search strategy programmatically applied, very much a different beast. Alas more spoof-able, and rather more sensitive to trends/happenstance. You'd have to admire them for trying that scheme out! :-)

That does not happen intra-run at E@H. Our strategy changes are made on a much longer time scale and with 'slow' humans explicitly in the loop. Committees, schedules, reviews, milestones etc. Very sedate by comparison.

Cheers, Mike.

You ain't just kiddin'!! I started looking into just what the !@#$ are they trying to do here, and ended up saying to myself, "Oh great, another graduate level self study course (or beyond) to plow through..."! :-)

Apparently they 'lost the handle' really quick a couple of times early on, but that might have had a lot to do with the really short preproduction work they were running. In my short time there so far, every new attempt seems to have gotten better or at least help showed the way for the next pilot run from the participants POV.

Then to have the 'Gamesters of Triskelion' arrive en masse and further complicate matters. What a joy! ;-)

Alinator

Winterknight
Winterknight
Joined: 4 Jun 05
Posts: 1450
Credit: 376874821
RAC: 141372

Having had a brief look at

Having had a brief look at that last link, why am I not surprised to see a certain character at the center of another controversy.

Comment viewing options

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