Pending credit: 164.90 granted O,O


log in

Advanced search

Message boards : Number crunching : Pending credit: 164.90 granted O,O

1 · 2 · Next
Author Message
Profile Guido_Waldenmeier_BiV
Send message
Joined: 3 Apr 99
Posts: 37
Credit: 0
RAC: 0
Germany
Message 161 - Posted: 9 Jun 2004, 18:27:48 UTC

[/url]
Long time ago far far away ;-)

Heffed
Volunteer tester
Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 164 - Posted: 9 Jun 2004, 19:03:27 UTC - in response to Message 161.
Last modified: 9 Jun 2004, 19:03:56 UTC

> [/url]
> Long time ago far far away ;-)

Remember Guido, this project wasn't even ready for the beta testers yet...

Profile Guido_Waldenmeier_BiV
Send message
Joined: 3 Apr 99
Posts: 37
Credit: 0
RAC: 0
Germany
Message 167 - Posted: 9 Jun 2004, 19:26:11 UTC

[/url]
i want my credits please MOM :-)))
Long time ago far far away ;-)

Profile MAOJC
Send message
Joined: 31 Jan 00
Posts: 11
Credit: 991,339
RAC: 0
United States
Message 177 - Posted: 10 Jun 2004, 1:17:58 UTC
Last modified: 10 Jun 2004, 1:18:52 UTC

You think you got a few stacked up! LMAO

Soon the results of the first round are trikling in!

John McLeod VII
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 15 Jul 99
Posts: 24785
Credit: 524,053
RAC: 86
United States
Message 180 - Posted: 10 Jun 2004, 2:06:52 UTC

Creditsare granted after three different hosts return results that the verifier determines are "the same". The credit granted will be the middle of the three credit requests.

This was done to promote science over cheating. It also makes certain that someone that has a problem with their machine will not be getting credit that they do not deserve. This includes machines that have been so severly over clocked that the FPU is not stable. Having an unstable FPU is not easy to diagnose, but it does show in S@H results. Apparently many results from S@H1 had problems with the science, either cheating of one sort or another, or trouble with the FPU.

With a small number of active hosts and a large number of WUs, it will take a while before each one is verified. With more active hosts, the time to have three valid results will be shorter.

There is a bug that has been noted that some of the WUs that have been crunched already have only required 2 results to validate. This is currently being looked into.
jm7

StieflerDuo
Volunteer tester
Send message
Joined: 19 Feb 03
Posts: 27
Credit: 17,330
RAC: 0
Germany
Message 185 - Posted: 10 Jun 2004, 7:13:57 UTC - in response to Message 180.

> Creditsare granted after three different hosts return results that the
> verifier determines are "the same".

I'd really like to hear that the results are verified now.
They weren't in Beta-Test and I don't think they are now.

> The credit granted will be the middle of the three credit requests.
>
> This was done to promote science over cheating. It also makes certain that
> someone that has a problem with their machine will not be getting credit that
> they do not deserve.

I think it was a good idea to raise the number of completed Workunits from 2 to 3 within the Beta-Test and cut the top and bottom claimed credit and grant the medium one. It should fit in most cases.

The only problem I see now is that if the results are not verified against each other, a "problematic host" will have lots of benefits from this.
In the Beta-Test we saw sometimes hosts that were producing an amazing lot of bogus results with a very low claimed credit within a small period of time.
The most annoying thing to the users was that if they returned the second result to the pair for these Workunits, they got the low claimed credit of the bogus host. Of course the people affected by this became angry about it, but at least the bogus hosts didn't get any real benefit from this because they only got the very small credits they've claimed, and even if they sent out lots of those bogus results the credit through the day didn't exceed normal limits.

Nowadays with the "grant the middle of three"-system these hosts would raise through sky's limit with their credit, because they'd get a reasonable granted credit for their impossible number of results per day.

My opinion is that as long as the server doesn't verify the results we didn't win anything.

> There is a bug that has been noted that some of the WUs that have been
> crunched already have only required 2 results to validate. This is currently
> being looked into.

We've seen this effect of "Didn't need"-Messages in the "Outcome"-row of Workunit's descriptions in the Beta-Test already.
This is being looked into for months now and I don't see any change.

Kind regards

Mike
[url=http://www.boinc.dk/index.php?page=user_statistics&userid=13991]

Profile MikeProject donor
Volunteer tester
Avatar
Send message
Joined: 17 Feb 01
Posts: 24881
Credit: 34,403,808
RAC: 12,709
Germany
Message 186 - Posted: 10 Jun 2004, 7:16:13 UTC

HI

But what i see is the same when Beta starts, much WUs with marked as unsent.
Which will never get credits granted.

greetz Mike

[url=http://www.boinc.dk/index.php?page=user_statistics&userid=9826]

Profile Shaktai
Volunteer tester
Avatar
Send message
Joined: 16 Jun 99
Posts: 211
Credit: 259,752
RAC: 0
United States
Message 189 - Posted: 10 Jun 2004, 7:46:04 UTC - in response to Message 186.

> HI
>
> But what i see is the same when Beta starts, much WUs with marked as unsent.
> Which will never get credits granted.
>
> greetz Mike
>

I don't think that is an error. It is due to the intentionally low number of participants. Those additional work units just haven't been sent out yet. Eventually they should be sent to other computers for validation.

Remember, this is still a test page. Only a handful of selected people are supposed to be testing it. Many others joined in too early, but not all work units have been assigned to 3 computers yet. Patience.


[url=http://team.macnn.com]

Heffed
Volunteer tester
Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 191 - Posted: 10 Jun 2004, 8:07:50 UTC - in response to Message 185.
Last modified: 10 Jun 2004, 8:10:56 UTC

> I'd really like to hear that the results are verified now.
> They weren't in Beta-Test and I don't think they are now.

I think you are forgetting that this is still beta. In fact, this project wasn't even intended for beta yet!

"Live" is still a ways off, contrary to what some people think. However, live was somewhat forced by some idiot that leaked the URL to this site.

If you take a look at the Main page, you will see that this project is indeed not live! :(

Here is the relevent text.
"June 8, 2004
NOTICE: this project is not officially active. Use at your own risk. We may delete results and reset credits at any time."


So please stop referring to beta in past tense. Beta is still going on regardless of how many testers jumped ship a little prematurely. (which hoses both this project, and beta)

Plus, now there are S@H1 people transferring over thinking it's "live", and they are now asking about all the "new" bugs that we're already aware of in beta. Not a fantastic "launch", that's for sure! :(

I think Dev should shut this site down until it's actually ready. All that can come of it is new users hopping in thinking it's the real deal, and getting discouraged by the problems. It has a net effect of making Dev, and the beta team look bad.

StieflerDuo
Volunteer tester
Send message
Joined: 19 Feb 03
Posts: 27
Credit: 17,330
RAC: 0
Germany
Message 195 - Posted: 10 Jun 2004, 9:02:03 UTC - in response to Message 191.

> I think you are forgetting that this is still beta. In fact, this project
> wasn't even intended for beta yet!

You're running through open doors here.
Trust me I'm better aware of this fact than you thought of.
It's me who asked others to stop sending people to this page without having a clue of what effects will raise from this.

But anyway, now we're here and when you here me referring to beta it's not like talking about 19th century but referring to the fact that we're discussing things already mentioned elsewhere, namely at Beta-Test.

In this special case I only pointed out that your thought of having a validator on the server-side yet might be wrong.

Kind regards

Mike
[url=http://www.boinc.dk/index.php?page=user_statistics&userid=13991]

Profile Guido_Waldenmeier_BiV
Send message
Joined: 3 Apr 99
Posts: 37
Credit: 0
RAC: 0
Germany
Message 197 - Posted: 10 Jun 2004, 9:05:47 UTC

sorry i write in german ,i can not good translate this what i will say
Leute
wenn diese theorie stimmen würde,das viele ergebnisse versendet werden aber wenige computer rechnen,wundert es mich das die ersten plätze so gnadenlos abziehen im vergleich zu den ersten 20 der liste unter
http://setiboinc.ssl.berkeley.edu/sah/top_users.php
und oh zufall die ersten 2 kommen aus den USA.
mir ist das system der drei sorten kredite schon klar im prinzip aber irgend wie ist da der wurm drin.also meiner meinung noch weit vom status eines beta testes entfernt.
nichts für ungut gruss guido

Long time ago far far away ;-)

Heffed
Volunteer tester
Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 202 - Posted: 10 Jun 2004, 9:50:14 UTC - in response to Message 195.
Last modified: 10 Jun 2004, 10:00:53 UTC

@ StieflerDuo

Sorry man, I'm just more than a little annoyed that someone mucked it up... :(

@ Guido

I tried a translator, but literal translations always come out a bit strange. It looks as though you are saying that since the top people are from the US, that someone in the US was the leak? That could very well be. I have no idea who might have done it, only that it was done.

I still feel Dev could find out who if they wanted to. Maybe cancel an account or something? Just a thought...

SirUlli
Volunteer tester
Send message
Joined: 3 Apr 99
Posts: 58
Credit: 28,048
RAC: 0
Germany
Message 205 - Posted: 10 Jun 2004, 10:01:13 UTC

@Guido,

look at the Fleet from MAOJC(Medium Array of Junk Computers)

http://setiboinc.ssl.berkeley.edu/sah/forum_thread.php?id=13

i see 24 Hosts

Greetings from Germany NRW
Ulli

Profile Guido_Waldenmeier_BiV
Send message
Joined: 3 Apr 99
Posts: 37
Credit: 0
RAC: 0
Germany
Message 210 - Posted: 10 Jun 2004, 11:56:31 UTC

that is true ulli looks like 24 *grinz*ich bin doch nicht blöd*

und kaufe bei MM ;-)
[/url]
Long time ago far far away ;-)

Profile eRazor
Volunteer tester
Send message
Joined: 30 Dec 99
Posts: 12
Credit: 62,805
RAC: 0
Canada
Message 218 - Posted: 10 Jun 2004, 15:02:10 UTC - in response to Message 191.

**Note: Not intended to be directed specifically at heffed. This is intended as an extension to the text below.

> > I'd really like to hear that the results are verified now.
> > They weren't in Beta-Test and I don't think they are now.
>
> I think you are forgetting that this is still beta. In fact, this project
> wasn't even intended for beta yet!
>
> "Live" is still a ways off, contrary to what some people think. However, live
> was somewhat forced by some idiot that leaked the URL to this site.

This project was intended to be a closed test of the data that was exported from classic sah, a bug hunting expedition for project "startup" problems, and database performance testing/tweaking with "production site" amounts of data. It was not intended to be public until the testing phase was complete, and credit reset. We don't even have a "production ready" core client yet (although it's very close).

Everyone can expect that the credits gained here will be zeroed before public launch. If you're in the game to amass credits, you'd be better off back in beta where you should be, and will eventually be rewarded in some fasion for your participation.


Profile eRazor
Volunteer tester
Send message
Joined: 30 Dec 99
Posts: 12
Credit: 62,805
RAC: 0
Canada
Message 219 - Posted: 10 Jun 2004, 15:10:18 UTC - in response to Message 202.
Last modified: 10 Jun 2004, 15:11:32 UTC

> I tried a translator, but literal translations always come out a bit strange.
> It looks as though you are saying that since the top people are from the US,
> that someone in the US was the leak? That could very well be. I have no idea
> who might have done it, only that it was done.
>
> I still feel Dev could find out who if they wanted to. Maybe cancel an account
> or something? Just a thought...

Have you ever tried to stuff a p1ssed off cat back into the bag? Even chainmail doesn't help :)

It's a done deal. It really doesn't matter where it came from so it's really fruitless to dig for the culprit. It just means that there will be more credits/results to zero out, and a lot more work for the team to sort through the bug reports for what's fact or fiction.

It's unfortunate though, that so many people have ditched the beta project where their participation is most useful and productive. Losing so many from beta may well delay the public launch.


Profile Guido_Waldenmeier_BiV
Send message
Joined: 3 Apr 99
Posts: 37
Credit: 0
RAC: 0
Germany
Message 222 - Posted: 10 Jun 2004, 15:21:15 UTC

[/url]
no comment no words
Long time ago far far away ;-)

Heffed
Volunteer tester
Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 227 - Posted: 10 Jun 2004, 19:47:44 UTC - in response to Message 222.

> no comment no words

Good call, Guido... :(

Profile Ageless
Avatar
Send message
Joined: 9 Jun 99
Posts: 12392
Credit: 2,665,317
RAC: 984
Netherlands
Message 229 - Posted: 10 Jun 2004, 20:28:02 UTC - in response to Message 180.

> Credits are granted after three different hosts return results that the
> verifier determines are "the same". The credit granted will be the middle of
> the three credit requests.
===============
Looks like it isn't working. ;)


----------------------
Jord™

SURVEYOR
Volunteer tester
Send message
Joined: 19 Oct 02
Posts: 375
Credit: 608,422
RAC: 0
United States
Message 232 - Posted: 10 Jun 2004, 20:46:10 UTC

Two have been completed and one is still in progress.

1 · 2 · Next

Message boards : Number crunching : Pending credit: 164.90 granted O,O

Copyright © 2014 University of California