Message boards : Number crunching : Credit Issued ?
Message board moderation

To post messages, you must log in.

AuthorMessage
KingPin

Send message
Joined: 2 Sep 04
Posts: 16
Credit: 26,713
RAC: 0
Message 2512 - Posted: 27 Sep 2004, 23:56:37 UTC

My Linux AMD boxes are being validated against Windows Intel boxes. Seeing that my credit on the Linux boxes are only half what they are on a Windows box. You guys with Windows boxes are getting lower credit when validated by a Linux box. Sorry. Check it out. :(:(

ID: 2512 · Report as offensive     Reply Quote
grumpy

Send message
Joined: 1 Sep 04
Posts: 57
Credit: 2,822,727
RAC: 412
Message 2517 - Posted: 28 Sep 2004, 0:18:12 UTC

Credit attribution here ( LHC) is, I think, Fuzzy logic anyway.

ID: 2517 · Report as offensive     Reply Quote
Profile Markku Degerholm

Send message
Joined: 3 Sep 04
Posts: 212
Credit: 4,545
RAC: 0
Message 2554 - Posted: 28 Sep 2004, 9:45:09 UTC

Our credit granting logic = BOINC credit granting logic.

Also, check Known Bugs.

Markku Degerholm
LHC@home Admin
ID: 2554 · Report as offensive     Reply Quote
Profile Paul D. Buck

Send message
Joined: 2 Sep 04
Posts: 545
Credit: 148,912
RAC: 0
Message 2609 - Posted: 28 Sep 2004, 20:13:00 UTC - in response to Message 2512.  

> My Linux AMD boxes are being validated against Windows Intel boxes. Seeing
> that my credit on the Linux boxes are only half what they are on a Windows
> box. You guys with Windows boxes are getting lower credit when validated by a
> Linux box. Sorry. Check it out. :(:(

There also indicators that the HT vs non-HT processors of the same speed give odd results. One of the reasons I stopped even looking ... I was just getting annoyed... so, no more Mr. Nice guy! :)

Seriously though, I think the project plan to get something that works and start beating one it is the right decision. Granted we pay for it with problems (what else is new in the world of computers?) and frustration. But, contrary to some, many of these bugs would have never been found even if we waited forever.

So, I think if we can get to a common code base with BOINC so we have ONE BOINC program, some of these errors should go away ...

Still, I am not sure how to really get the arms around something like this, you would have to have identical hardware, then process the WU on it under all OS (well, the Mac is going to screw up part of that plan maybe) ...
<p>
For BOINC Documentaion: Click Me!


ID: 2609 · Report as offensive     Reply Quote
grumpy

Send message
Joined: 1 Sep 04
Posts: 57
Credit: 2,822,727
RAC: 412
Message 2652 - Posted: 29 Sep 2004, 1:04:46 UTC - in response to Message 2554.  

> Our credit granting logic = BOINC credit granting logic.
>
> Also, check Known
> Bugs
.
>
> Markku Degerholm
> LHC@home Admin
>

Our credit granting logic = BOINC credit granting logic. ====>>> ????????

Attribution credits ===>>> VALIDATOR ====>>>> PROBLEMS >>>> Floating Point math+ differents CPU+OS+FORTRAN+ ...???


Reference: Your FAQ page article 2.4

2.4 How and when do I get credit?

The credit system is described here. For each workunit three results are created.
At least two will need to be valid before credit is granted. If the third one returns valid after the two first have been validated,
it will also be granted credit.

Please note that because current validation system requires identical results, credit will not be granted in the following situation:
-your host gives some result (possibly correct)
-two other hosts give two identical results which differ from yours

Currently at least old Pentiums, new Pentiums and new Athlons seem to give different results.
So getting credit from good job is somewhat based on luck. We are working to improve the situation.


DO YOU CATCH THE WORD L U C K. ( YOU WROTE IT, NOT ME).

LHC NEWS

20.09.2004 15:30 UTC
*****Major Upgrade*****

New Sixtrack version with better performance and more pleasing graphics.
Among other things the memory leaks have (hopefully) been fixed -
Please test this!

New and more relaxed validator has been installed, which should take care of validation problems.


YOU CONTROL THE VALIDATOR AND IT'S ALGORITHMS THUS ATTRIBUTION LOGIC.
(You wrote it, not me)



DEFENITION: ( for those who don't know)

Fuzzy Logic is a superset of Boolean logic dealing with the concept of partial truth.
Whereas classical logic holds that everything can be expressed in binary terms (0 or 1, black or white, yes or no),
fuzzy logic replaces Boolean truth values with degrees of truth which are very similar to probabilities
(except that they need not sum to one). This allows for values between 0 and 1, shades of gray, and maybe;
it allows partial membership in a set. It is highly related to fuzzy sets and possibility theory.
It was introduced in the 1960s by Dr. Lotfi Zadeh of UC Berkeley.

The AND, OR, NOT operators of boolean logic exist in fuzzy logic, usually defined as the minimum, maximum, and complement;
when they are defined this way, the are called the Zadeh operators, because they were originally defined as such
in Zadeh's original papers. There are also other operators, more linguistic in nature, called hedges that can be applied.
These are generally adverbs such as "very", or "somewhat", which modify the meaning of a set using a mathematical formula.

Fuzzy logic has suffered many misconceptions, partly due to its name. "Fuzzy" is said to have negative connotations,
usually either suggesting something cute or something imprecise; the latter sometimes causes people to
equate "fuzzy logic" with "imprecise logic". However, fuzzy logic is not any less precise than any other form of logic,
rather it is an organized and mathematical method of handling inherently uncertain concepts; the concept of "coldness"
cannot be expressed in an equation (temperature is a quantity, but "coldness" is not). However, everybody has an
idea of what "cold" is, and agrees that something cannot be "cold" at N degrees but "not cold" at N+1 degrees
(which is a concept classical logic and equations cannot easily handle).
Another common misconception is that fuzzy logic is a new way of expressing probability.
However, Bart Kosko has shown that probability is a subset of fuzzy logic, as probability only handles one kind of uncertainty.
He has also proved a theorem that shows that Bayes' theorem can be derived from the concept of fuzzy subsethood.
This should not by any means suggest that all those who study probability accept or even understand fuzzy logic,
however; to many, fuzzy logic is still a curiosity.

Fuzzy logic is also sometimes said to be used only in AI, control systems, and/or expert systems (note that these fields
can have significant overlap). These are by far the most common applications, but by no means the only possible; fuzzy logic
can be applied in any situation requiring the handling of uncertainty.


golden_goose



ID: 2652 · Report as offensive     Reply Quote
Profile Chrulle

Send message
Joined: 27 Jul 04
Posts: 182
Credit: 1,880
RAC: 0
Message 2718 - Posted: 29 Sep 2004, 13:12:29 UTC - in response to Message 2652.  

> YOU CONTROL THE VALIDATOR AND IT'S ALGORITHMS THUS ATTRIBUTION LOGIC.
Validation logic != credit assign logic

Validation simply decides wether a result is considered good. Here we are different from the other projects.

But once it is decided to grant credits to a result the credits are calculated and given according to exactly the same rules as Boinc/Seti.


Christian S�ttrup
LHC@home Developer
ID: 2718 · Report as offensive     Reply Quote

Message boards : Number crunching : Credit Issued ?


©2024 CERN