Can some one tell me how someone with only has 3,822.20 points get a RAC of 30,128.86??

6 PF2002 30,128.86 3,822.20 Netherlands 3 Jan 2006 15:17:02 UTC

I have 66+ PC's running 24/7 and upload 2X's his over all total every 6 HR's and my RAC is 31,650.90 Thats only 1500 more then his

Is this guy cheating by changing the numbers on his end

Do you have his computer ID? If so please contact me at joseantonio@choicecable.net or contact BOK. Do so ASAP, please

No his Computers are hidden from the public

---

If You Want The Best You Must forget The Rest
---------------And Join Free-DC----------------

A lot of the first time posters are showing strange results like that, Lauren2.

Credit: 4
RAC: 250
single or double digit Credit, and RAC in the 3 digit range. Is RAC being calculated differently with Boinc 5.4.9 or Rosetta's .22 release?

This is not 250 BUT 30,128 His RAC is 8X his points And if you look all his points are way over 30 days old. So I am not sure 5.4.9 was inplay then

You're looking at large numbers, and I'm looking at smaller numbers - but the result is the same.
People are getting rediculously high RAC (? Average Credit) scores for tiny amounts of credits lately. It's like they're making their first upload, and the RAC algorithm is assuming they'd only worked for a short period of time, and multiplying to get a daily score. Look at their RAC a few days or uploads later, and it should be less than their credit.

---

@Laurenu2: You are one of those, having a BOINC version 5.5.0 with far too high benchmark results. What kind of BOINC client are you using? Why are the benchmarks that high?

Can you tell me what computer node are you referring to that is to high.
I even get lost in my network 70+ nodes Please give computer name not computer ID #
I have been restocking with new AMD X2's, are these the ones your talking about
If you look at the Top participants My first listing is about 20+ pages down
and the people in front of my first listing are using lesser computers then I.
So I would guess I am not getting that much more points for the work done

But as I said to the system Lords it would be best to make it all even for the same work done

---

If You Want The Best You Must forget The Rest
---------------And Join Free-DC----------------

LOL...

I am running 2 Proc Servers and I am getting credits of 25 to 30 each. Must be doing something really wrong.

:-)

---

@Laurenu2: You are one of those, having a BOINC version 5.5.0 with far too high benchmark results. What kind of BOINC client are you using? Why are the benchmarks that high?

Can you tell me what computer node are you referring to that is to high.
I even get lost in my network 70+ nodes Please give computer name not computer ID #
I have been restocking with new AMD X2's, are these the ones your talking about
If you look at the Top participants My first listing is about 20+ pages down
and the people in front of my first listing are using lesser computers then I.
So I would guess I am not getting that much more points for the work done

But as I said to the system Lords it would be best to make it all even for the same work done

Rac is not calculated by the Project. It is calculated by all the stats sites. However you can artificially raise the RAC by extending the time between results reporting to the system. In other words, assuming two identical systems, one reporting in every few hours, and the other reporting in one a week. The system that reports results once a week will show a higher RAC.

The Rosetta project is doing what it can in dealing with any credit issues. Since the calculation of RAC is not part of the project, that will have to be done somewhere else.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

@Laurenu2: You are one of those, having a BOINC version 5.5.0 with far too high benchmark results. What kind of BOINC client are you using? Why are the benchmarks that high?

Can you tell me what computer node are you referring to that is to high.
I even get lost in my network 70+ nodes Please give computer name not computer ID #

https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=221521

This one has quite high credits-per-minute, between 3 and 5 times what I'd have expected a CPU of this type to get (particularly when you consider that it's doing two work units at a time). Benchmark is 12.5K, about 3 times higher than my single-core A64 overclockedclocked to 2.5MHz.

Owner Laurenu2
Created 16 May 2006 6:23:41 UTC
Total Credit 45,384.05
Recent average credit 1,521.67
CPU type AuthenticAMD
AMD Athlon(tm)64 X2 Dual Core Processor 3800+
Number of CPUs 2
Operating System Microsoft Windows XP
Professional Edition, Service Pack 2, (05.01.2600.00)
Memory 447.29 MB
Cache 976.56 KB
Measured floating point speed 3829.02 million ops/sec
Measured integer speed 12506.96 million ops/sec
Average upload rate 6.88 KB/sec
Average download rate Unknown
Average turnaround time 1.07 days
Maximum daily WU quota per CPU 100/day
Results 201

Typically it's getting 65 credit per 6,500 second WU (and obviously will be processing two at a time, hence 1 credit per 50 seconds.

---

This is sounding more and more like the SETI boards before the big brouhaha erupted. So let's make something clear: Hell is going to freeze over before we allow the SETI situation to happen here. Need I remind people how a comment insinuating that a person was cheating degenerated in a mess and ugly name calling.

So to all of you posting here : if you are going to question some-one's production DO IT IN PRIVATE to them . This is not the place to question a person production.

Let's make this perfectly clear: the use of optimized clients is not perse , cheating. First of all their existence is well known to Boinc users and second their use has not been prohibited by the developers of Rosetta. So, please. stop using cheating and optimized clients in the same sentence. Doing so is an invitation for a flame war that we do not need in here and we don't want in here.

Lets also make clear that many of the suggestions proposed to create an "even field benchmark" are not feasible for Rosetta as the multiple quorums required is a waste of computing resources.

Be conscious that the developers are working on an internal benchmark appropriate for the scope of this project and the type of crunchers that join this project. Let us let them work on it, test it and submit it to the comments and suggestions of the members of this community. Do not make their work harder by situations like the one that is starting to develop here: specific people are having their production ( and BTW their credibility ) questioned. It took very little to have the big brouhaha at SETI started, I repeat don't start in here.

Right now there is a more important issue than "fixing" the credit issue ( if there is really a credit issue); it is called CASP 7 let's not loose sight of this. Anything that diverts our energies from CASP 7 is not helpful to the project.


BTW: don't think that the developers are not looking at credit claims that look unreal and cannot be explained rationally: they are working on it as we speak. Please let the group that is working on that do their work without having to become fire fighters putting down flames and flamers.


A final question: does one really believe there is a single credit system that will keep everyone happy? So why do we keep asking for what we know is but a philosophical impossibility?

This is not SETI, do not bring your SETI baggage and conflicts in here. Enough is enough.

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

Can some one tell me how someone with only has 3,822.20 points get a RAC of 30,128.86??

6 PF2002 30,128.86 3,822.20 Netherlands 3 Jan 2006 15:17:02 UTC

I have 66+ PC's running 24/7 and upload 2X's his over all total every 6 HR's and my RAC is 31,650.90 Thats only 1500 more then his

Is this guy cheating by changing the numbers on his end

Do you have his computer ID? If so please contact me at joseantonio@choicecable.net or contact BOK. Do so ASAP, please

No his Computers are hidden from the public

There is the posibility a person merged all his ghosts into a single real computer/host.

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

I really did not want to offend or attack someone. If someone felt attacked or offended by me, please excuse me.

cu, Jochen

[Edit:] I just tried to edit my posts, but I can't anymore... Is there a moderator that could delete them?

---

While 3rd party stats sites may use their own algorithms to create a daily average credit rating (that makes more sense than the Boinc RAC does to most of us.. :) the Rosetta server is creating the Credit score and RAC that are being listed at the left of our messages - unless I've been misled.
Laurenu2 mentioned seeing a system with a RAC 8 times larger than its actual credits; and I've seen similar results (much smaller, though) from some of the first time posters lately.

For one of the project programmers - Is there a problem with the RAC calculation and the new 5.50 server code? Are the systems with the larger RAC than actual credit scores from ghosted systems - i.e. the problem BAM users? i.e. the Boinc RAC calculator can't handle ghosted systems?

---

I really did not want to offend or attack someone. If someone felt attacked or offended by me, please excuse me.

cu, Jochen

[Edit:] I just tried to edit my posts, but I can't anymore... Is there a moderator that could delete them?

Jochen,

Yes, I can do that. But which ones do you want deleted?

As a personal aside (Jose need not read further),

The Rosetta forums are open for discussion of any topic, so long as the discourse is polite and professional. While I see no need for my intervention as a Moderator in this discussion beyond explanations as necessary, this topic has really been discussed to death.

There are 70,000 people participating in Rosetta. Very few have raised this as an important issue. But from my point of view it is counter productive to keep raising the issue in the context of the thinly veiled threat that masses of people will leave if it isn't fixed right now. The project team is aware that a very few systems may be claiming higher credits than might be justified by the system type. They are working on a credit system that does NOT rely on information that might be altered by a user (or the users BOINC client). When this is in place there will be an announcement.

In the meantime it is unreasonable to expect the project to drop everything and spend a lot of the limited programming resources available to specifically address this problem ahead of other issues. Let me be clear about this. They are already working on the issue, but it is NOT going to be handled ahead of the CASP work which is being conducted against a very unforgiving set of deadlines. The new credit system will be integrated into RALPH as new versions are prepared for testing.

I along with a few others were asked by the project to take a look at some of the credit claims made by some systems, and recommend a solution. The fast solution would be to simply delete the users who "seem" to have questionable credit claims and purge the credits from the system. I think you can all agree that would not be the best solution. It is arbitrary, and open to a lot of subjectivity. So that means something must be done that would require a more surgical approach. But that also means it will take more time.

For every credit granting solution I have seen proposed from the user community, I have seen an objection raised, from another part of the user community. In other words, one persons solution is another persons poison. At present, the Flop counting method is the most fair proposal on the table, and some people do not like that either. No matter what credit system is used, there will be someone who thinks it is unfair.

The Rosetta project team, working in concert with their counterparts among the BOINC developers, have a working plan they are implementing. But it will require coding and testing on RALPH before it will be deployed in Rosetta. That will take time. How much time is anyones guess, it depends on a lot of factors. But no matter what that schedule works out to be, arguing over it on the forums will not influence or speed up the process. Nor will threatening to leave the project if it is not fixed fast enough.

The fact is that of the 149,000 systems currently shown as attached to the project less than 70 have been singled out by the user community as problematic. That represents a very, very small percentage. While it could be argued that there are more than 70 and they just have not been reported, the current evidence does not support that position.

There is a larger set of systems that fall into a grey area where they may be claiming higher than normal credits, but not so high as to be completely impossible to justify. With these systems there may be any number of reasons why the claims are in fact legitimate. I for one am not prepared to enter into a witch hunt and demand that these system owners submit justifications for their credit claims.

What is forgotten in all this is that there are in fact a lot more systems that actually claim LESS credit than they should. Either the benchmarks are off, or there was some other process that reduced the benchmark when it was run. Also many Linux systems claim LESS credit than they should just because they are running Linux. The project should also provide universal EXTRA credit to those systems as a matter of fair-play, but I don't see many of you arguing for that.

So my points are these. The project is working on the problem. It is not as widespread as these discussions would have people believe. The immediate solutions available to the project will not make everyone happy. Even the long term solutions will not make everyone happy. While I frequently see people say that "The science comes first...", that is almost universally followed by a comment like "if you don't fix the credits, I and a lot of people will leave".

This project is doing work that WILL save lives, maybe even yours. How long should the project team shut down the real work to fix the credit system? Either the science comes first or it does not. You can't have it both ways.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

Typically it's getting 65 credit per 6,500 second WU (and obviously will be processing two at a time, hence 1 credit per 50 seconds.

Which would result in 36 credits per hour per CPU.
My FX-60 gets 357 credits for a 24 hour work unit: 15 credits per hour per CPU.

Just a wild stab in the dark: You are using some optimized clients for other projects and therefor run a calibrating BOINC application and forgot to turn off calibration for Rosetta?

cu, Jochen

Jochen
Can I ask how old your FX-60 system IS ?
Do you use the system at all? Do you have any firewall, anti Virus running?
Do you use that FX-60 for anything except crunching Rosetta?

This computer is brand new and has both cores working ONLY Rose
The OS is NEW with no programs loaded except Rosetta and There is no I/O like sound, mouse/KB to rob cycles
And yes I do have them OC But Look At there OUTPUT 100jobs per week
That machine is 2 weeks old and did 200+ jobs
your FX60 has been working for 6 months and has only done 21 jobs
With a record like that I can see why your grumpy with High RAC's
You might want to do some general maintenance defrag. spy ware blow out all the cooling, reset the CPU with new paste.
But if you look at all the finished jobs I did I am sure you will see I do run faster cleaner PC then most just by the number of jobs I complete compared to other members with the same PC in the same time frame

---

If You Want The Best You Must forget The Rest
---------------And Join Free-DC----------------

Jochen,

Yes, I can do that. But which ones do you want deleted?

Actually all my above posts.

Dear Moderator9,

First of all, I'd like to apologize for warming up a topic that has been discussed to death.
I wanted to know, if I'm doing something wrong. I was looking for a fault in my installations.
After reading Jose post and my posts again, I agreed with Jose and that is why I asked to remove my posts. They could cause disturbance. That is something I don't want.

This project is doing work that WILL save lives...

This is the most important reason, why I don't want to cause any further trouble.

Thank you for your detailed explanation. I will try to be more careful in the future.

Kind regards,

Jochen

@Laurenu2: As recommended, we should continue this discussion in private. I'd be happy to hear from you. My e-mail adress is: jr <at> jr101.de

---

Dear Moderator9,

First of all, I'd like to apologize for warming up a topic that has been discussed to death.
I wanted to know, if I'm doing something wrong. I was looking for a fault in my installations.
After reading Jose post and my posts again, I agreed with Jose and that is why I asked to remove my posts. They could cause disturbance. That is something I don't want.

This is the most important reason, why I don't want to cause any further trouble.

Thank you for your detailed explanation. I will try to be more careful in the future.

Kind regards,

Jochen

@Laurenu2: As recommended, we should continue this discussion in private. I'd be happy to hear from you. My e-mail adress is: jr <at> jr101.de

Jochen,

Please do not misunderstand my post. I was not upset with you about this thread. But I will do as you asked.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

[quote]As a personal aside (Jose need not read further)[quote]

As one of the people that is working on on the documentation intensive job of identifying those "weird credit claims" I get frustrated to see what is going on. If I sound harsh and I will plead guilty as not to be diplomatic it is because I know of the efforts by the Rosetta Developers to workout and develop a fair benchmark. But I cannot expect nor will expect the scientists to drop everything to deal with such a minor issue as credits. That is not their primary job.

So people know: one of the problems we, who are working with the questioned claims , is that there are cases we are dealing with unique machines, very radical machines so modified to the point of extracting all the crunching power possible. In many cases machines that only run Rosetta. At least in my case I am trying to document every single machine ( to the extent of my abilities) and comparing the claims against the benchmark databases that are available. It takes time, specially when for reasons that are obvious I will submit many of the computers I have documented to another person for follow up and for check. My team members can vouch I have even asked them for more details of their machines. And right now, I and other people are running some trial runs of possible programs that could affect benchmarks to get a better comparison. This takes time specially when one wants to be fair and when one doesn't want to be disruptive of the science project that is going on.


I have been very honest with those who asked me to join the task of checking the credits about my fears of how this credit thing can degenerate in an useless and counter productive brouhaha like the one that happened in another community. I do not want that for Rosetta.

I have also been very honest on the fact that I do believe that credits are but the frosting of a nice cake. And that as such they have a good use: I have seen the search for credits used to motivate people that have computers capable of producing a lot of models to enter the fray. To me that is the use of credits...motivation to get computers in: nothing less , nothing more.

That said: If I have offended someone with my candor and my bluntness: please accept my apoogies. To the Moderators my apologies if I in the way I express my opinion have not helped to keep the civility that has been the norm here.


That said: please if any of you has a question regarding claims, feel free to contact me at joseantonio@choicecable.net.

Jose

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

I find it hard to believe that Rosetta or Boinc can make the point system in place now a fair and even field, way to Manny clients, OS's, and type CPUs , conch's, And Hyper Overclocking to deal with.

Back to my original Question about RAC How can a member get a Rac that is

8X his total points

I might be slow but I am having trouble understanding how the RAC is formulated

I know it is not of any real value here. But I do like to understand how things work

---

If You Want The Best You Must forget The Rest
---------------And Join Free-DC----------------

I find it hard to believe that Rosetta or Boinc can make the point system in place now a fair and even field, way to Manny clients, OS's, and type CPUs , conch's, And Hyper Overclocking to deal with.

Back to my original Question about RAC How can a member get a Rac that is

8X his total points

I might be slow but I am having trouble understanding how the RAC is formulated

I know it is not of any real value here. But I do like to understand how things work

Well, this link provides the formula by which it is calculated. I am certain the WIKI also has more detail about this. But the short of it is the number does not mean anything. It only gives a basic measure of a contribution over time. The reason that very low scores produce very high RACs is because the time element of the formula does not balance out in the short term properly. For the formula to work properly you need a number of samples taken over a longer period of time. Remember this is a "average". the more samples used in the calculation, the more accurate the average will be.

Think of it this way. If the RAC is calculated for a single day, and a person submits one result for 25 credits in the first hour, you would logically project that that person should produce 600 credits over the 24 hour period based on the submission in the first hour of the day. Over time the falsehood of the assumption dissolves when it turns out the system actually only produces that single result in a 24 hour period. Eventually that RAC would become 25 over time.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

Thanks Mod9 that helped a little I think
Computer code always make my eyes go Cross eyed

---

If You Want The Best You Must forget The Rest
---------------And Join Free-DC----------------

Thanks Mod9 that helped a little I think
Computer code always make my eyes go Cross eyed

In my case code drives me to drink :)

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

Thanks Mod9 that helped a little I think
Computer code always make my eyes go Cross eyed

In my case code drives me to drink :)

I always wrote my best code when drunk. The comments didn't make much sense but the code worked !! Hick-up... Hick-up.. :)
Thinking about it.... thats probably the reason I am now in Hardware... lol