Best of luck to anyone writing to David about CreditNew.
Maybe if he gets a couple hundred emails he realises it's not as good as he must belive it to be.
There's always hope.

Just because I've made abundantly clear that I couldn't care less for credits, doesn't mean that I don't believe the current credit system to be inherently flawed and in dire need of fixing (for several reasons).

If you decide to do a thing, do it right.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Juan, I think the part of the message from Dr Anderson that you should read is this:

I've never received email from the people who threaten to write letters to the chancellor; they should try doing that.

(Emphasis added.)
In other words, Dr Anderson is inviting such people to contact him directly.

I allready send a e-mail to DR. A. in the past, by the lack of even a simple aknowledge of the e-mail i belive is easy to reach the center of a black hole...

Even with this, i belive the "via chancellor" way is not a "polite" way tho try to fix the issue. But that is only a personal opinion, I´m not sugesting anyone to do nothing or not to do nothing, please don´t missunderstood my words... That way follows a "dangerous path", some kind of "last resort like nukes".

I agree with william, maybe a 100´s of e-mails could help, if he don´t trate them as SPAM...or a DOS attack...

BTW I don´t blame creditnew itself, what i ask for is only balance the credit of AP vs MB, it´s simple... it´s easy, just adjust the base line of any one of them and leave the rest untoched. IE base of MB now is 100 change to 200, or AP is 600 change to 300, so they will paid proportionality aproximately the same amount of credit for both WU, like was before V7, nothing else. I not ask for more credit or for changing in the statistical analyis. Maybe that is what nobody could understand.

Ops... my mistake, i do it again... we all need to stop, "we a wasting the bandwidht" please forgive-me, shut down for a another minute to compensate that "loss".

---

Same as it ever was,
same as it ever was.

Hi folks I have a large mug of calm-the-heck-down(aka coffee) and will happily share.

A bit of info just for reference..
Dr. Anderson does not like imperfections to be pointed out.

The staff are doing what they can with what they have to work with.

Along the way things WILL get balanced out again. Slowly. We are talking Glacial. Pre climate change.

The staff will not come in here and tongue lash you for not thinking everything is great. They know it is not. But they would really rather have a cup of coffee.

We all see some value to looking for E.T. . Not all of us see the same value.

Cheering and Jeering each other is not going to help matters. A grand sense of self importance is really going to end up embarrassing. For everyone. So have some coffee. It really is good.

If you must write a letter, then write a letter. Just be ready for less than satisfactory responses. And no, I do not want to know what they are.

In the mean time, I recommend crunching what you can get, let the credit situation work its way through the mill, and coffee. Lots and lots of coffee.

Same as it ever was, same as it ever was.

---

Janice

Same as it ever was, same as it ever was.

You might think...

Meanwhile, Dr A is in the news over this side of the pond.

---

Hi folks I have a large mug of calm-the-heck-down(aka coffee) and will happily share.

Decaffeinated I hope ...

---

... In the mean time, I recommend crunching what you can get, let the credit situation work its way through the mill, and coffee. Lots and lots of coffee.

Same as it ever was, same as it ever was.

There are various solutions to make the 'credits' more reliable, consistent, and scientific. Myself and at least one other have mentioned one calibrated solution a few times over the years...

However...

When it comes down to it, the credits are seen as being nothing more than a "bit of fun" and for myself, I've never felt any need to start coding anything better. The cobblers are already imperfectly keeping some amused.

So... Please feel free to program your own solution for a new Boinc credit system. I can recommend the well thrashed out system of NIST-style hierarchical calibration and reward. There's various descriptions for how that can work in threads of old on these forums.

Sorry, but I feel no urge to develop and promote the credits improvements/code myself. Quite reasonably, I'm sure Dr A will have better things to do with his time also.


Instead, sponsor a summertime student to come up with a fix?

Happy fun ever-faster crunchin'!
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

There seem to be a lot of misconceptions about the relationship between CreditNew and the decrease in credit granted for some people.

First, we've been using CreditNew for far longer than S@H v7 has been around. CreditNew is not all that new.

Second, there are no knobs to turn to increase credit. That was kind of the point of CreditNew. It more accurate to say that all the knobs are still there, it's just that turning them is counteracted by the CreditNew mechanism. I know, I've tried to turn them all with no effect.

Third, the credit decrease is by no means universal. For S@H v6, the elapsed time based credit for the stock cpu app on the average windows machine was scaled by 0.984311268359135. For S@H v7, the elapsed time based credit is scaled by 0.997756082254495. That's right, according to our server, windows machines are getting 1.3% more credit per second than they did under v6.

Where might the difference people are seeing come from? Maybe GPU apps. The efficiency of GPU based apps is different on between v6 and v7. Some parts of the code are more efficient, especially if you've got a GPU optimized for CUDA5. But the balance between different portions of the code has been changed by autocorrelation. Since we grant credit based on actual work, your GPU might be getting more or less credits per second. I'm not sure there is anything I can do about that. I tried raising credit by boosting the GPU projected flops multiplier by a factor of two. It worked for about a day before the GPU flops scale adjusted itself down by a factor of two.

And then there are optimized apps. We can't make any guarantees about the rate at which a non-stock app gets credit. Since credits are referenced to the stock apps, if we optimize the stock app, the credit rate for non-stock apps will go down. There was lots of wailing and gnashing of teeth when we released SETI@home enhanced, because optimized apps that used to get 10X the credit per second of the stock app suddenly only got 3X the credit. Lot's of people threatened to quit the project, and some did. Again, this is nothing I can fix.

As for other projects the grant more credit... Maybe its time to ask them why. For many projects there seems to be no relationship between credit granted and work done, especially GPU-only projects. CUDA5 GPUs are about 2.5% efficient when running SETI@home, and that's actually not that bad for GPU code. If we assumed GPUs were 100% efficient, and had a GPU-only app we could boost your credit rate by 40X. But I prefer that they mean something.

---

@SETIEric@qoto.org (Mastodon)

Thank you for the explanation Eric.

---

Thank you Eric.

---

[/quote]

Old James

Thanks very much Eric for taking time and trouble to post in this forum about the CreditNew.
Really appreciated.

---

Eric, thanks for the details in your posting, I actually think I understand it! I know, that's a scary thought...

John

---

Third, the credit decrease is by no means universal. For S@H v6, the elapsed time based credit for the stock cpu app on the average windows machine was scaled by 0.984311268359135. For S@H v7, the elapsed time based credit is scaled by 0.997756082254495. That's right, according to our server, windows machines are getting 1.3% more credit per second than they did under v6.

Thanks Eric for all that information.

Just one thing left: how does that look like compared to Astropulse? Are stock astropulse applications getting also more credit per time than stock MB? Or is all that discussion about astropulse getting much more credit per time just because no one here ever tried stock applications?

---

Third, the credit decrease is by no means universal. For S@H v6, the elapsed time based credit for the stock cpu app on the average windows machine was scaled by 0.984311268359135. For S@H v7, the elapsed time based credit is scaled by 0.997756082254495. That's right, according to our server, windows machines are getting 1.3% more credit per second than they did under v6.

Thanks Eric for all that information.

Just one thing left: how does that look like compared to Astropulse? Are stock astropulse applications getting also more credit per time than stock MB? Or is all that discussion about astropulse getting much more credit per time just because no one here ever tried stock applications?

I think that this is the main issue by some as doing AP's now gets a person much more credit than MB's by fair distance presently (resulting in some doing stupid things just to get more AP's).

But I'll keep trying to keep the MB's rolling.

Cheers.

And then there are optimized apps. We can't make any guarantees about the rate at which a non-stock app gets credit. Since credits are referenced to the stock apps, if we optimize the stock app, the credit rate for non-stock apps will go down.

This point (as I wrote few times already) makes CreditNew approach countereffective for its ultimate goal regarding this particular project - to INCREASE peoples involvement and participation in project!
BOINC's staff can use "credits" as some measurement tool - it's ok per se (and requires some precision increase that CreditNew tries to achieve), but for this particular project it's not such tool, it's additional stimul for some of participants to donate resources for very this project. And CreditNew fails in this MOST IMPORTANT goal of "credits".
IMO it should be rejected.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Lot's of people threatened to quit the project, and some did. Again, this is nothing I can fix.

I don't think so. At least you could initiate discussion with BOINC staff about goal of "credits" per se. For what reason BOINC projects have such entity at all???

As for other projects the grant more credit... Maybe its time to ask them why.

There is at least one and quite simple explanation: because they want to use "credits" in its initial meaning and initial goal: to STIMULATE participants for donating their resourses. Not as some measurement tool that allows other peoples to make some investigations and conclusions based on this number....
(bibliometric obsession close resembles situation with credits btw ;) ).

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Third, the credit decrease is by no means universal. For S@H v6, the elapsed time based credit for the stock cpu app on the average windows machine was scaled by 0.984311268359135. For S@H v7, the elapsed time based credit is scaled by 0.997756082254495. That's right, according to our server, windows machines are getting 1.3% more credit per second than they did under v6.

Thanks Eric for all that information.

Just one thing left: how does that look like compared to Astropulse? Are stock astropulse applications getting also more credit per time than stock MB? Or is all that discussion about astropulse getting much more credit per time just because no one here ever tried stock applications?

Try to run stock CPU app and answer will be obvious...

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Try to run stock CPU app and answer will be obvious...

Sure I could do that, but the numbers I get would only be valid for my particular hardware and that might be quite a bit off of the mean values for "average CPU", which Eric has. Those numbers are a lot better than any of us can get by trying the stock apps. And if the stock CPU AP and MB pay about the same on "average CPU", than the issue here is not CreditNew but less optimized stock astropulse app, which makes astropulse paying a lot better for people running optimized applications.

The only issue left with CreditNew would be than the fluctuations in granted credit for the same type of WU, but that's completely different story.

---

than the issue here is not CreditNew but less optimized stock astropulse app, which makes astropulse paying a lot better for people running optimized applications.

Exactly (about less optimised stock) that you would see running it on own hardware and on any another hardware too.
But I would disagree that it's not CreditNew issue. Instead of rewarding for stock app bettering it punishes already installed opt custom apps if stock becomes better! Is it fair way to go, really ? I think it's very bad way to go.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

But I would disagree that it's not CreditNew issue. Instead of rewarding for stock app bettering it punishes already installed opt custom apps if stock becomes better! Is it fair way to go, really ? I think it's very bad way to go.

I think it's the correct way to go. Everything anyone get above the stock level can be considered as a bonus for installing the opt apps. So now the bonus will be a bit smaller and if the fully optimized app became stock one day, we would basically all run stock and get stock credits.

I don't see any reason why the credit for stock should be raised if the app becomes better optimized. If you work somewhere and your employer buys you a new tool, which will make you work more efficiently so you get more work done per day, I highly doubt he's gonna raise your salary for that.

---

If you work somewhere and your employer buys you a new tool, which will make you work more efficiently so you get more work done per day, I highly doubt he's gonna raise your salary for that.

And if you were working for someone, and you supplied your own tools which made you more efficient & you earned bonuses for that work, you'd be pretty p'd off if the business then got those tools for everyone else, and did away with the bounses.

---

Grant
Darwin NT