Easy solution: Credits = Time spent on it X benchmarks (or something that measures the PCs performance)

If people get optimized versions.... then the time will decrease, but you'll get more done in the same time, so you'll have about the same credit (ppl, don't whine if u get 10 credits less, this is ABOUT SCIENCE, NOT CREDITS)

Lets stop this strike stuff and lets crunch like good ol' times :)

---

I just want to thank daniel and anyone else involved and dedicated to bringing these issues forward. I agree with all the requests and am a participant in the demonstration. I believe the bug issues will be worked out soon,since i cannot see why they would be allowed to continue so issues #2,3 and 4 are my main focus!

---

Easy solution: Credits = Time spent on it X benchmarks (or something that measures the PCs performance)

If people get optimized versions.... then the time will decrease, but you'll get more done in the same time, so you'll have about the same credit (ppl, don't whine if u get 10 credits less, this is ABOUT SCIENCE, NOT CREDITS)

Lets stop this strike stuff and lets crunch like good ol' times :)

Agreed it is about the science!! (the main reason why I participate)

But please keep in mind that some people measure their participation/accomplishments by their credits earned. Since the release of enhanced some feel as if their earnings/contributions has been devalued. (note: I am making this assumption) for those whom credit does matter simply would like the developers to reconsider the possiblity of an adjustment.

@Ned
Sorry to ask Ned but....why not give one credit per WU then instead of having a system that actually makes most people unhappy? I'm sure cheating could be intercepted in other ways now that the project is so much wiser.

Sorry, I'm not Ned, but I would like to address this issue.
With the extremely variable nature of the enhanced work units (etc.)

The other question, though, is exactly which definition of "fair" do people mean when they say "we want fair credits!"

I picked the one I think fits the context -- "fair" seems to mean "equitable."

But that may not be what they're asking -- another definition says "adequate."

[edit]My goal here is to clarify the request: if they want "adequate" and the project delivers "equitable" that won't resolve the issue.[/edit]

Hmmm sorry to come back.
Fair or equitable - give 1 credit per WU on the basis everyone is likely to get a predictable and equal distribution of "long" and "short" units. Berkrley could even attempt to manage that a fair distribution is secured. Result = you lose out sometimes with long units but you gain others with short units and end in balance. Just like everyone else. If you have faster hardware your reward is to get more credits for doing more work. If you optimise then you get the additional benefit of more units. I think the whole "credit system failure" thing is impacting far too hard BOINC'S credibility. Sadly I might add.

The problem with this is that it is all to easy easy to look at the Angle Range of the work unit before processing it and simply delete the ones with high ARs - those that generally take longer. Done with care (not too agressively), the quota system won't catch a host/user 'dropping' WUs that they don't want to run.

This isn't theory, it was common practice with classic SETI and the programmming to write a 'dumper' program for high AR WUs is trivial.

Hey, if someone could shuffle those high AR WUs (1.12 and up) my way, I'd take every one of them. On my A64, using Crunch3r's 5.12, I'm back to 4.18 processing times (~2500 seconds) and the highest credit per hour rate (18-22) I have. They're the best WUs for me and from the data I've collected, they're best for just about everyone.


kev
data

---

kev

X2 4400+,4200+ @2.75GHz, XP1800+ @1.65GHz, P4 @1.6GHz

Well stated Daniel!

---

This is only a test...

Sorry for my late reply. The mountain has been in lightning shutdown, and we're about to get hit again.

1.) Crunch3r, who has been instrumental in developing optimized applications for Seti and Seti Enhanced, was accused of cheating by some of the forum members.

I don't know what we can do about this. We have no control of forum members aside from shutting down the forums entirely. If things don't settle down soon, we may be forced to do so.

and had also sent Eric Korpela an email with a list of C++ errors in the standard Seti Enhanced application, that were eliminated when he built version of the Enhanced application WITHOUT optimization.

Look at the message. There is no list of C++ error messages aside from the alloca() problem which is specific to the Intel compiler. He claims the C++ exceptions are eliminated without giving any clues to what caused the exceptions in the first place. He doesn't even mention which of the numerous types of exceptions (which are Windows exceptions, rather than C++ exceptions) this claims to fix Given that I have some understanding of what is causing some of the exceptions (and having recently solved some of them if I get the new client into beta.), I can tell you that using a different compiler is highly unlikely to solve any of them.

If Crunch3r claims otherwise then he should pick one exception type, show how to reproduce it with a stock client, and then show how it does not occur with his recompiled client.

Eric had denied receiving this email, and even went so far as to call Crunch3r a LIAR. Crunch3r sent this email to Eric TWICE, and has proof that Eric received the second email.

Yes, and Crunch3r and others have called me worse things than LIAR. I'm not sure what this "proof" is. I had not seen the message until I explicitly looked at my second level spam folder and subsequently let it through. That was well after Crunch3r "quit" the project.

Thus, Crunch3r has left the project, and is no longer working to develop Seti code any further. This alienation and vilification of this project's top contributors needs to stop.

And I think the alienation and vilification of the Project Scientist needs to stop. You think Crunch3r's departure is hurting the project? My departure would end it, not that I intend to leave. But I don't need to sit here and be attacked. I have certainly faced more attacks in the last few days than Crunch3r ever got.

We request that Eric issue this apology in a new "Sticky" thread, and that the thread be reserved for a dialog between Eric and Crunch3r to work out their differences.

Thank you for acknowledging that I have appologized. Like many involved in this flame war, I got angry at being attacked and lost my temper for a few minutes, and I am sorry. But I don't see any apologies from people directed toward me...

Until I get some indication from Crunch3r (his email address will now get through my spam filter. I've added it manually) that he has some regrets about how he has acted and some indication that he wishes to take part in discussion, a dialogue is not going to happen.

Please remember that Crunch3r is taking this action in part because he didn't receive an answer to one email that he had sent me. He did not attempt to contact me again to see if I had received it. Pardon me if I see that as an overreaction.

I also don't think I would see anything but personal attacks by others in such a forum. As part of my response "requests," I'd like to see the unwarrented attacks on my character by others in the forums stop. Negotiation is a two way street...

I think I am owed apologies from a large number of people who have posted in the forums. I don't expect to get them...

2.) The application errors in the code are causing countless hours of wasted CPU cycles. The application crashes out or gets -9 overflow errors MANY hours into workunit processing. This results in no credit being granted at ALL for such
workunits. Thus, we request an application that does not waste our CPU cycles by going into a seemingly endless loop.

That problem was a server side problem that resulted in invalid thresholds in the workunit header. It has been solved. Incidentally, like the other bugs recently fixed, Crunch3r's recompiled application would not have affected this one way or another.

3.) There have been many discussions on what is "Fair Credit" for the work we perform. Since credit IS important to a large number of participants, we request that this situation be investigated sooner, rather than later. Perhaps looking into a system that takes into consideration the TOTAL amount of work done, and not just the FPOPS.

Essentially all of the work done by SETI@home is the FPOPS, so it would be difficult to come up with a better means of measurement. As a result of recent measurements I've increased the credit by 5% in a version that will be going into beta this week. An increase by significantly more than that is unwarranted at this time.

There are some misconceptions that the benchmarks in the core client are an indication of the amount of real work that an application can do. There is also a misconception that the amount of work done must be exactly proportional to time. Some areas of the code will always do more work per CPU cycle than others. The benchmarks are essentially an upper limit to the amount of work that can be done. The original idea in BOINC that the benchmarks would measure the total work actually done was misguided, and I said so at the time that idea was being developed.

One of the main reasons people are unhappy with the new credit system is that the original SETI@home was so suboptimal that people were able to build optimized versions that had 8 times the performance. That meant people could get their machines to generate 8 times the credit per hour by using an optimized version. Now that the better optimized enhanced client has been released, a factor of 8 is no longer possible. That has made some people angry. There is no change to the credit system I can make that will appease them.

4.) Being as that "some" of the forum moderators are doing more to feed the flames rather than to help extinguish them, we request perhaps a complete sweep of the forum mods with all mods elected by forum members.

I will discuss this with the rest of the team when I get back to Berkeley. I did not choose the moderators, nor do I wish to take part in choosing moderators. But I think a voting scheme will likely lead to the same problems we already have, lots of four letter adjectives flying around saying why so and so shouldn't be a moderator. Maybe it's time to go to a zero tolerance moderation scheme where any attack by anyone on anyone will be immediately deleted. There seems to be a widely based "It's OK for me to attack you but you can't attack me" attitude. I've certainly seen it directed at me.

Apparently a lot of people feel they can use any four letter adjective to describe me but that any reciprocation on my part is out of bounds. I think we all need to tone it down. I hope we can refrain from swearing and name calling in the forums in the future.

---

@SETIEric@qoto.org (Mastodon)

Ned, please read #3 again and try to understand. All we are asking for is that they look at it again "sooner" and not "later" and reconsider what they say is fair. We are not trying to define it for them but are asking them to consider the issues about it that were already brought up and possibly look at other ways to determine fair credit.

I'm not Ned (and I don't play him on TV either), but you have not really defined what #3 means. For example, what is the difference between "the TOTAL amount of work done" vs. "just the FPOPS" in your mind? Obviously you perceive there to be a difference. Bear in mind that this is scientific work. Floating Point calculations are done quite often in scientific work.

Brian,

Thanks for giving this thought. You're on the right track but my question is a little broader.

The on-line definition of "fair" is here.

I'm pretty confident that Sarge doesn't mean definition #1 (pleasing to the eye)

I think #6a fits the best (marked by impartiality and honesty : free from self-interest, prejudice, or favoritism).

... but I suspect that some people think they deserve more credit (definition #10).

So, that is why I started asking -- because when I thought about "fair" I realized that the word, like so many simple words, has more than one meaning.

-- Ned

---

Look this Thread posting was a position taken By Eric K. to try and stop the disputes here and not continue to increase the flames.

So everyone try and be more conciliatory in your postings please.

---

Come and Visit Us at
BBR TeamStarFire


****My 9th year of Seti****A Founding Member of the Original Seti Team Starfire at Broadband Reports.com ****

I for one appreciate the long hours and tireless dedication to the project, Eric.

Your response echoes many of our opinions at BBR Team Starfire. While Starfire hasn't endorsed the strike, some of our members have switched to other projects until the bugs get worked out, a few in support of the strike, some actually ramped up production because of the it.

We feel the credit system is fair, even to the detriment of Starfire as a ranked team. My personal comparisons to the Rosetta credit on identical rigs shows Seti credit a bit low to it. What more could be asked for is roughly the same credit between BOINC projects?

I thank you for your time, and am sorry you've had to deal with this. I've yet to donate to the project directly, but will soon do just that, I've been inspired....

edited for clarity...

Paul

---

Team Starfire World BOINC
IRC- irc//irc.teamstarfire.net:6667/team_starfire

I would like to thank you and your team for all of your hard work. I for one, simply divided time on my machines between S@H and E@H to assure that they always have work to do. Enhanced SETI is new and everyone should expect issues early on. I'm sure once the dust settles everyone will calm down in regards to the new credit system. I do believe it is fair or very close to it.

I hope everyone will have the patience while the bugs get worked out.

---

Eric,

Thanks for taking the time out of your busy schedule to address these issues. I appreciate all the time and effort you and the whole Seti@Home team has put in over the years, and I'm personally sorry to see the attacks that have been made on you and the project team.

The personal attacks and accusations are just plain wrong on all sides, and unfortunately this board has degraded largely into what ammounts to a gang fight. I think a zero tolerance policy with moderators who only moderate, and don't fan the flames, is a great idea.

You're willing to step back and look at the situation objectively, and I think everyone should be willing to do the same. Alot of feelings have been hurt on all sides, and it's time to just let that go and move forward with continuing to make seti@home the best DC project out there. What's happened here over the past weeks can do nothing but scare people away from the project.

---

XaaK

Very good responses Eric.......

I have used the optimized apps from Crunch3r for a long time. I have alway's understood that my RAC was artificially inflated because of this. I have seen my RAC drop from 6500 to it's present level and I have not and do not complain.

Thanks for your hard work and dedication to this project Eric.

---

Boinc....Boinc....Boinc....Boinc....

Did the topic of this thread change?

It seems odd that people who supposedly have no interest in the Strike are sure quick to come here and offer their anti strike .02....They won't support the strike but they won't pass up any opportunity to put it down either.

It's not like their feelings arent widely known from their postings in many of the OTHER more appropiate threads on this topic to make these types of comments in....

Why don't both sides return to neutral corners? Leave this thread for those who support the work stoppage and maybe start seperate thread(s) defending Berkeley's way of doing things and stop trying to demean what's happening in this thread..

---

Consider me foolish..

But requests were made. And now we have received a response.

I thought that a discussion based on the response follows? But if one whom is against what we are attempting would certainly tell us now "you received a resonse..now put it to rest." That seems a bit premature. I however am confident that if we discuss this in a civilized manner, the "dust will settle" and Seti will be a better place...

Did the topic of this thread change?

It seems odd that people who supposedly have no interest in the Strike are sure quick to come here and offer their anti strike .02....They won't support the strike but they won't pass up any opportunity to put it down either.

It's not like their feelings arent widely known from their postings in many of the OTHER more appropiate threads on this topic to make these types of comments in....

Why don't both sides return to neutral corners? Leave this thread for those who support the work stoppage and maybe start seperate thread(s) defending Berkeley's way of doing things and stop trying to demean what's happening in this thread..

I second that.

---

http://www.setiusa.net

Did the topic of this thread change?

I see Eric posted a very detailed and awesome message that I think everyone should read. Was it off topic? I don't think so.

It seems odd that people who supposedly have no interest in the Strike are sure quick to come here and offer their anti strike .02....They won't support the strike but they won't pass up any opportunity to put it down either.

Would you expect anything different? Everyone goes off topic. THe way I see if, you wanna strike, then strike, if not, then don't. I don't think its gonna hurt anyone's feelings either way.

It's not like their feelings arent widely known from their postings in many of the OTHER more appropiate threads on this topic to make these types of comments in....

I agree.

Why don't both sides return to neutral corners? Leave this thread for those who support the work stoppage and maybe start seperate thread(s) defending Berkeley's way of doing things and stop trying to demean what's happening in this thread..

Asking this would be like asking everyone to stop posting for a week. I don't wanna sound rude or off here, but it won't happen. You know it, I know it, and so do others.

As Eric stated, everyone needs to cool off. Personally, I would not blame the folks who administer these forums for taking it completely down. Do they want to? No. They know this is where people with problems and idea's can come and share them and help each other out. I'd like to see this occur again. Maybe one day, when this all smooths over, but that won't happen soon.

---

Consider me foolish..

But requests were made. And now we have received a response.

I thought that a discussion based on the response follows? But if one whom is against what we are attempting would certainly tell us now "you received a resonse..now put it to rest." That seems a bit premature. I however am confident that if we discuss this in a civilized manner, the "dust will settle" and Seti will be a better place...

We may need to change the venue for the discussion with Eric.

---

http://www.setiusa.net

It sounds like you just don't like responses that disagree with your view.

---

-----

Doh.........I didn't mean to pull the plug on your strike! Please continue the strike. While your at it some here need to learn what a strike is.

Traditionally those who go on strike have to give something up to join the strike. That usually means that you give up your pay check and put your bills and family at risk! Many here think it is fine to just set "no network access" and continue crunching for the week. What are they giving up?? That action is NOT a strike that is delaying the normal paycheck.

Anyway........please continue the strike.................if that's what you call it.

---

Boinc....Boinc....Boinc....Boinc....

I see Eric posted a very detailed and awesome message that I think everyone should read. Was it off topic? I don't think so.

I agree that Eric answered on a point by point basis what was asked of him, that's not what I was refering to when I asked "Did the topic of this thread change"?

I was making an observation that since this was a thread for the 'pro strike' people, it's odd (actually it's really not) that the 'anti-strike' crowd would come here and offer their .02 why they don't feel like they need to support this..

I say to that, let them go start their own thread and return this one to it's original intended purpose which was to open a dialog with Eric Korpela, Rom Walton, Dr. David Anderson, etc... WHOMEVER it needs to be opened/started with to effect change based on the issues objectivly presented here at the beginning of this thread..

Otherwise this thread will degenerate in to just another flamefest and be summarily ignored by the very people it was meant for...