I've just switched back to Rosetta@Home from WCG because of the unfairness with credit on Linux systems. I'm more for the science of course, but since Rosetta@Home is still partly based around the HIV/AIDS virus, I'll be running R@H until WCG get their fixed credit system in place.

And even if both projects were equally fair, here each credit is new science. On projects running redundancy only half (or less) of the credits are science, the other half (or more) being used to check the answers.

This holds for the old and new credit systems here, of course, so technically I am off topic...

There is on major thing witht he new credit system,
It stops people like Jose having to search through the credits looking for client_state file manipulators and stop things like this over at XtremLabs (which the project seems totaly unaware I guess)
http://xw01.lri.fr:4320/top_hosts.php
Loads of Top Hosts using file manipulation (general use 'optimised' clients would never claim that high). Since XtremLab have a max credit/hr that's not a problem, just set you Pentium4, D Even the Pentium 3 and 2's in the list you'll see to just under that.

That now has little effect with the Rosetta@Home's 'new' credit system.

---

Team mauisun.org

There is on major thing witht he new credit system,
It stops people like Jose having to search through the credits looking for client_state file manipulators

Well yes, if the new system spoils Jose's fun that might count as a disadvantage... ;)

But actually it is still possible to inflate host stats on the new system - Run several identical hosts for a while. Detach / Re-attach all but one. Wait for the detached hosts to have all their results deleted, leaving only their credits, merge with the one "master" box.

Repeat every so often.

Of course, it is different from client_state manipulators in two important ways - work of that amount of credit has actually been done, all that is tricky is the assigning of it to one host. And secondly, although it unfairly raises a box in the host stats, it does not have an unfair effect on user/team stats, about which most users seem more concerned.

The serious point in the above is that not all of those accused of being client_state editors were doing what was suspected. No doubt some of them were.

R~~

There is on major thing witht he new credit system,
It stops people like Jose having to search through the credits looking for client_state file manipulators

Well yes, if the new system spoils Jose's fun that might count as a disadvantage... ;)

But actually it is still possible to inflate host stats on the new system - Run several identical hosts for a while. Detach / Re-attach all but one. Wait for the detached hosts to have all their results deleted, leaving only their credits, merge with the one "master" box.

Repeat every so often.

Of course, it is different from client_state manipulators in two important ways - work of that amount of credit has actually been done, all that is tricky is the assigning of it to one host. And secondly, although it unfairly raises a box in the host stats, it does not have an unfair effect on user/team stats, about which most users seem more concerned.

The serious point in the above is that not all of those accused of being client_state editors were doing what was suspected. No doubt some of them were.

R~~

lol :-D

Did you have a look at how many of the computer over at XtremLabs I went throuhg 3 pages and all where claiming near the maximum the project allows. As far as I know it's mainly the two top people there. I know one of them from Boinc@Hull and he certainly is and he's doing it because the other person is from Boinc@Australia.

---

Team mauisun.org

...

Well yes, if the new system spoils Jose's fun that might count as a disadvantage... ;)
...

R~~

Can I ask a favor? Do not mention my name.

The way you are stating it: I hunted for cheaters for my temporal enjoyment. I did not. I did so, because I believed that the real cheaters here harming the atmosphere here. Alas those who spouted baseless accusation did more harm than them.

I have never enjoyed hunting for real cheaters.

Every time I documented a cheating episode in my life ( as a Compliance Auditor, a teacher, an uncle and yes, as a volunteer here) a part of me died:

I do believe in the basic honesty of people. I was raised that a person only takes to his grave his reputation for honesty and for his capacity to love so, finding dishonest people affected me a lot and affects me: my belief in the intrinsic good of people and their honesty dies a little and I am worst for it.

That is why I get very angry when a baseless accusation of dishonesty is made. It goes against the basic nature of people.

Oh since I have been here I have noticed that David Baker mentioned he needs about 40 Tera Flops of computing power to process the proteins related with Alzheimer's , and that many of you are basically stating that work units need to be segregated by the computing powers of the machines ....It seems that many of the smaller, less powerful crunchers cannot handle the very large Wu's that are coming down the pipe.

So it seems that the asseveration made by many of you that:
1- The big powerful computers could be substituted by the new people that were going to barge in as soon as the credit systems was fixed was going to be more than enough to continue.
2- That the computing power of certain teams was not needed.

was plain and unadulterated wishful thinking by many of you.

:)

---

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

Well said Jose!

If I understand correctly ...

It is a shame that people will cheat the system for the sake of credits. Inflating your stats or your host's stats projects a false 'impression'. At the end of the day you go to sleep knowing for a fact, you and/or your host is not as good as it seems.

Being at the top of any list means zilch if you have not contributed to the science or did the work to deserve it. There is absolutely no fun or any value in falsely claiming any reward.

How can anybody believe he is the best if it is not so? ... I just don't get it.

---

I like to watch my credit totals rise, makes me feel like I'm doing something useful, but I don't dwell on it. I got curious today and calculated the credit per hour that I get for Rosetta and Einstein (the only 2 projects that I indulge). This isn't scientific by any stretch since I only looked at the most recent 6 results at each project. Using first grade derivation and integration I got 9.67 credits per hour at Einstein and 6.58 credits per hour at Rosetta. Therefore I received 47% more credit per hour at Einstein than Rosetta.

If I was more credit conscious I would dedicate my machine entirely to Einstein. A few weeks ago I increased my resource share to Rosetta and since then my world position has been steadily dropping. Purely for selfish reasons I think Rosetta's granted credits are a little low relative to other projects?

Slight change of topic here. I was just looking at my recently returned results and noticed that I am granted more credit than I claim on every result returned. Now, under the new credit system, I realise that this is not unexpected but surely I should have some where I am granted less?

I am interested if this implies anything about my machine.

Any thoughts anyone?

Stwato

---

Slight change of topic here. I was just looking at my recently returned results and noticed that I am granted more credit than I claim on every result returned. Now, under the new credit system, I realise that this is not unexpected but surely I should have some where I am granted less?

I am interested if this implies anything about my machine.

Any thoughts anyone?

Stwato

It implies you machines are just crunching faster than boinc's credit system would have thought it did.
And looking at your CPU, a Pentium-M 1.6GHz it is scoring correctly.
A lot is probably down to the fact you have [like my Pentium-M 1.7GHz (o/c to 1.9GHz)] a 2MB cache and rosetta likes that :-)

---

Team mauisun.org

Slight change of topic here. I was just looking at my recently returned results and noticed that I am granted more credit than I claim on every result returned. Now, under the new credit system, I realise that this is not unexpected but surely I should have some where I am granted less?

I am interested if this implies anything about my machine.

Any thoughts anyone?

Stwato

Not necessarily - I get on average around 2.5x granted/claimed credits - that's because I'm running on a Linux platform, where the BOINC benchmark is somewhat poorer than the Windows machines.

Also, if your processor is "more efficient than average" then you're going to get more credit than those with a lower than average processor efficency.

Since you have a Pentium M, you have a slightly faster machine than the Pentium 4 users, who are the majority of Rosetta crunchers.

--
Mats

---

Excellent!! Thanks for the answers guys, it's good to know that I'm doing more work for the project than I thought. I hadn't realised that Rosetta likes cache so much.
All very interesting and quite exiting if I am

"more efficient than average"

.

Crunching on...

Stwato

---

So, I'll shoot myself in the foot, and as where the preferred thread/location for reporting credit calculation problems is?

This result, at 214 claimed/900 granted seems a bit out of character for one of the machines working for my team. The three previous results for this machine were: 214/280, 211/265, 210/256.

Either the machine got lucky and was result #2 for a work unit where result #1 came from somebody's basement supercomputer, or perhaps there is some problem with credit adjustment?

---

It looks like you are getting twice the number of jobs (8 six hour jobs per day) for which you get half the credit of comparable machines.

Look at this host. It has the same specs and gets about 100 credits per 6 hour job (might be slightly overclocked).

I'm not following. For the 27th, I returned 7 jobs (I was just finishing transitioning from 8 to 6 hour jobs). I looked at the machine you referenced. It had 8 jobs on the 27th too. Seems to me they pretty much mach with regard to number of jobs.

Besides the difference in credits, I am seeing no other difference....except for the OS of course.

I checked the processes. There are only the two rosetta application running, each at about 100%, which is right for a dual core processor. For what it's worth, it looks identical to the ps output from the Pentium D machines.

---

It looks like you are getting twice the number of jobs (8 six hour jobs per day) for which you get half the credit of comparable machines.

Look at this host. It has the same specs and gets about 100 credits per 6 hour job (might be slightly overclocked).

I'm not following. For the 27th, I returned 7 jobs (I was just finishing transitioning from 8 to 6 hour jobs). I looked at the machine you referenced. It had 8 jobs on the 27th too. Seems to me they pretty much mach with regard to number of jobs.

Besides the difference in credits, I am seeing no other difference....except for the OS of course.

I checked the processes. There are only the two rosetta application running, each at about 100%, which is right for a dual core processor. For what it's worth, it looks identical to the ps output from the Pentium D machines.

Try to set this host to a new venue, let's say "school" and set number of processors to use to 1 so that only one instance is running. Then let it crunch 24 hours and see how much credit you receive for 6hour WU. I'm sure we will narrow this down in the end.

The work claimed versus work granted must always be a disincentive for MAC users, also the variation in the percentage granted causes concern in either the way the percentage is calculated or in the way work is distributed or in the way MAC's are utilised.
claimed ? Granted granted %
11 Jun 2007 15:42:58 UTC Over Success Done 12,338.63 31.57 22.97 72.76
11 Jun 2007 18:37:17 UTC Over Success Done 9,615.18 24.60 12.06 49.02
11 Jun 2007 14:44:07 UTC Over Success Done 8,280.15 21.18 7.08 33.43
11 Jun 2007 13:08:47 UTC Over Success Done 10,629.90 27.19 11.70 43.03
11 Jun 2007 9:43:43 UTC Over Success Done 9,758.72 24.97 10.53 42.17
11 Jun 2007 6:42:34 UTC Over Success Done 10,667.17 27.29 13.81 50.60
11 Jun 2007 3:29:26 UTC Over Success Done 10,631.33 27.20 9.50 34.93
11 Jun 2007 0:17:44 UTC Over Success Done 13,686.35 35.01 23.12 66.04
Totals 85,607.43 219.01 110.77 50.58

---

The Rosetta code isn't ideally suited to the PPC architecture unfortunately. It was being looked into but then Apple announced they were switching to Intel CPUs so the priority for this dropped (coding time is a limited resource as ever).

The is variation in the credits as an effect of the credit system, although it works quite well generally. I don't know if PPC cpus show more variation in the scores though - maybe they do better on some work units than others???

Danny

---

Contrary to what is said about there not being a bad side to credits....I say you must have been born blind. Crunching a wu for 6 hours and I receive 10 credits? I should have received the claimed credit of 47. Credit granted should coincide with the amount of time that a computer crunches a work unit and NOT some arbitrary solution dreamed up to satisfy someones itch. It is plain to see so don't offer up any explanation that would otherwise say so. When these wu's are done crunching then it seems it will be time to get rid of Rosetta altogether. People crunch not only for science but for credit. It's a natural innate action to try and get as much credit as possible. When you begin granting appropriate credit for work done then I may return....if it's worth it.

Am I to understand that you are advocating the same amount of credit be granted for 6 hours of crunching to both an Intel core i7 960 architecture at 3200MHz and a Pentium 3 architecture at 766MHz?

How about if that Pentium 3 766MHz was also performing another cpu-intensive task (streaming/converting video/audio files, perhaps) during those same 6 hours that Rosetta was crunching, and the core i7 was a dedicated cruncher, with no other out-of-the-ordinary tasks on the cpu?

If not, then what "arbitrary solution" would you "dream up" ?

Scientific work actually completed, perhaps?

Credit granted should coincide with the amount of time that a computer crunches a work unit and NOT some arbitrary solution dreamed up to satisfy someones itch.

<Maxwell Smart voice> Sorry about that, Chief. </Maxwell Smart voice>

It is plain to see so don't offer up any explanation that would otherwise say so.

Sorry about that too.

When these wu's are done crunching then it seems it will be time to get rid of Rosetta altogether.

True. And for extreme cases, there's a slang term, not invented by me, that's bandied about: "credit whores".

People crunch not only for science but for credit.

References from scientific peer-reviewed journals, please.

It's a natural innate action to try and get as much credit as possible.

If not, as an alternative, you can expend the financial, intellectual and hardware resources necessary to create your own DC project, and feel free to create your own credit system, as you see fit.

Perhaps you'll want to grant 100,000,000 credits per hour of crunching, regardless of the amount of work completed?

And since "It's a natural innate action to try and get as much credit as possible", ever single Boinc cruncher (presently 1.6 million) will stop crunching for every other Boinc project (presently about 60), and every single Boinc cruncher, all 1.6 million of them, will then crunch for your project.

And then, all will be right with the world, correct?

When you begin granting appropriate credit for work done then I may return....if it's worth it.

duplicate post.

Contrary to what is said about there not being a bad side to credits....I say you must have been born blind. Crunching a wu for 6 hours and I receive 10 credits? I should have received the claimed credit of 47. Credit granted should coincide with the amount of time that a computer crunches a work unit and NOT some arbitrary solution dreamed up to satisfy someones itch. It is plain to see so don't offer up any explanation that would otherwise say so. When these wu's are done crunching then it seems it will be time to get rid of Rosetta altogether. People crunch not only for science but for credit. It's a natural innate action to try and get as much credit as possible. When you begin granting appropriate credit for work done then I may return....if it's worth it.

Credit granting has always been a touchy issue for each project. Whatever they do they always get compared to some other project that is doing it differently. Personally I would like to see the credits wiped out every January 1st and we start over for the new year. A baseball players stats, as an example, do not carry over year to year. Sure there are total stats and I am not advocating not keeping a record of a users total stats. But each user should start over every year, this would tend to settle down the "natural innate action to try and get as much credit as possible". Credits are a rank of how many computers each of us have access too. I see my stats are blowing you out of the water, does that make me a better person, or even a better cruncher? NO IT DOES NOT!!! It means I have access to more pc's than you do, nothing more. It means I am able to contribute more to each project than you are able to, but it doesn't mean you are no less valuable to that same project! Stats are for the user, not the project. Stats have had to be carefully monitored because some people have tried to scam the system and get credits they were not entitled to.

The granting of credits, and how much to grant, has caused untold hours of discussions at very high levels and still there is no consensus about how much to grant and exactly how to do it. The newest thought process is what Bad Penguin has said, credit granted for the amount of Scientific work actually done by each of our pc's. So if you have a 400mhz pc and I have a 4ghz pc, both crunch for 6 hours and each doing nothing else, I would get more credits than you do, simply because my pc would complete more work than yours does. Does this demean your pc, no it just means that each of us has to get faster and faster pc's to keep up with the Jones' of this World that ARE getting faster and faster pcs'. In the end it is the projects that benefit the most, they get more work done in less time.

---

Credits are a rank of how many computers each of us have access too. I see my stats are blowing you out of the water, does that make me a better person, or even a better cruncher? NO IT DOES NOT!!!

That is wrong. It says that you contributed more to this project than somebody else.