I think, it should be banned.
Nothing to do with science, just money.
Comments?

600 users making 90% of ALL BOINC CREDITS????
How can this be right?

Shrug. It's not as if the credits are worth anything, so until they do, it doesn't really matter to the powers that be.

---

Shrug. It's not as if the credits are worth anything, so until they do, it doesn't really matter to the powers that be.

Agreed. I've never seen anything where it's been said Boinc had to be **ONLY** about Science-related projects.

BURP isn't a science project either, I hope you don't want to ban them as well. I think a more reasonable request would be to have the project look at their credit distribution, is it as skewed for non-ASIC devices?

---

My Detailed BOINC Stats

Shrug. It's not as if the credits are worth anything, so until they do, it doesn't really matter to the powers that be.

There's one place where the BOINC credits do mean something scientific (or at least they're supposed to): the front page of this website.

There's a little box in the top right-hand corner labelled 'Computing Power', and underneath it says (currently) "24-hour average: 6.990 PetaFLOPS".

BOINC doesn't record or pass out statistics on how many floating point operations have been performed, so how does that get there? Answer - it's reverse engineered, calculated from the total user RAC at each project, on the assumption that every credit is worth the same.

When Bitcoin Utopia first went ballistic, that 24-hour average reached 17.063 PetaFLOPs - which, given that Bitcoin ASICs perform integer operations only, was getting perilously close to scientific fraud. At least some common sense has prevailed now.

I think a more reasonable request would be to have the project look at their credit distribution, is it as skewed for non-ASIC devices?

Maybe, but then that's normally done by the users rather than the BOINC development team, for the simple reason of what to do when they would request that a project follow their credit sampling, and the project says "no"?

Not much there can be done to force a project to follow the will of BOINC development, while users can en-masse say that they leave the place. Aside from perhaps not being included in the Projects List. Everything is open source, everything can be hacked, edited, reprogrammed, adjusted, tinkered with to the hearts content of the project administrator.

Now to give you an idea: even you can start a project today that pays twelve trillion credit per task done on the Uppercase application included in the package. There have been fan-based projects in the past, there are probably still fan-based projects this moment, that run nothing but Uppercase and give out credit for that. All of those might be diluting the value Richard points out, I don't know.

---

All of those might be diluting the value Richard points out, I don't know.

Well, since I'm helping to look at various aspects of the credit system at the moment, I thought I'd have a scout around.

I looked at three of the major statistics sites: BOINCstats (which tends to be regarded as authoritative by David Anderson); BOINC Combined Statistics (the in-house stats site - the power behind those 'Projects in which you are participating' lists on your account pages at BOINC projects); and Free-DC (the source of the graphs on BOINC's Top 100 multi-project BOINC participants page).

The three sites all collect and aggregate their data at different times of the day, and monitor a different number of projects (between 67 and 145). At the time I collected the figures, the grand total computing power was reported to be between 13 PetaFlops and 32 PetaFlops on the different sites, using that RAC --> GFlops reverse calculation.

But all of that was wildly different interpretations of the Bitcoin Utopia phenomenon, which the three sites were displaying as 6.2, 18.1, and 25.3 PetaFlops respectively. Strip out that one project (which seems to be accounting for 80% of all credit awarded, these days), and the residual figure is a much more consistent 6.8 PetaFlops across all three sites - and pretty close to the 6.99 PetaFlops we started with here.

Looking in particular at the 'Top 100' page, I thing we can identify visually that Collatz, DRTG and PrimeGrid are other high-paying integer math projects (I'm not sure why DRTG shows so strongly - it seems to have been offline for the last couple of months). Other high-paying projects are SETI, Einstein, and GPUGrid. SETI and Einstein have large numbers of computers attached: GPUGrid certainly does a lot of floating point maths, but I think rather overstate the work done by their credit awards. I could see no sign of any large-scale input from the sort of Uppercase project that Jord describes.

If we leave out those three big integer projects, then BOINC's overall floating-point computing power comes out at 4.0 - 4.2 PetaFlops. That still puts it (if the methodology stands up) on a respectable par with ~#11 on www.top500.org - but no way is Bitcoin Utopia the second-most powerful supercomputer in the world, as its daily credit output would have you believe.

but no way is Bitcoin Utopia the second-most powerful supercomputer in the world

It might not be the second-most powerful general purpose, programmable supercomputer in the world, but those little single-purpose ASICs are orders of magnitude better than any CPU or GPU for the intended task ;) Much like Colossus - Would have been a terrible performer for general tasks, but it was pretty good at cracking enigma codes :)

Cheers,

Al.

Nothing to do with science, just money.

I think you might be forgetting what BOINC stands for. It's the Berkely Open Infrastructure for Network Computing. It's a software developed by Berkely University for volunteer computing over the internet. If somebody wants to volunteer their computer for a non-science project, that's their choice.

600 users making 90% of ALL BOINC CREDITS????
How can this be right?

BOINC credits don't really mean anything. It's just a statistic. Different projects give out different amounts of credit for tasks completed. You don't really get anything from it; if anything it's just a way of saying how much work you've done.

If somebody wants to volunteer their computer for a non-science project, that's their choice.

I think it's a good idea to make a sort of "ethics commetees" to evaluate the projects and to ban, if necessary.
What do you think if an Army uses BOINC to simulate new weapons??

Shrug. It's not as if the credits are worth anything, so until they do, it doesn't really matter to the powers that be.

Funny, how it went from "measuring" contributed flops to meaning nothing :D

Anyways, as Retvari Zoltan says on gpugrid:
http://www.gpugrid.net/forum_thread.php?id=3796&nowrap=true#37766

What if someone starts a BOINC project, which does the following:
1. gives 130 million BOINC credits for each USD its users donate for the project
2. keeps the 12% of the donated money for itself
3. sends the rest of the money for projects which apply for it
4. decreases the amount of BOINC credits awarded per USD on a regular basis

and also
http://www.gpugrid.net/forum_thread.php?id=3796&nowrap=true#37782

What is the purpose of the BitCoin utopia project? To generate and give funding to projects. (I do appreciate that.) It is a research? No. If it's not a research, then no BOINC credits should be awarded. BOINC credits shouldn't be awarded especially in exchange of money donations, as every project has a special place for their money donors on their homepages, still none of them gives BOINC credits for money. I my opinion that would be a shame if they would. Now the BitCoin Utopia project made that shame routine. Would it be okay, if the GPUGrid project gave Hypernova 426.8 billion credits for his 2500EUR donation? Does the BOINC community want to change its moral that way?

I would like to propose myself for handling ur money, plz plz plz donate
/With Luv, kind stranger
(I keep 25%, have many cats to feed)

Nothing to do with science, just money.

I do no know think they are breaking any rule or anything along those lines but I do believe that with the credit granted they do appear to have quite a great value. and perhaps there should be a different set of stats for them.

600 users making 90% of ALL BOINC CREDITS????
How can this be right?

This is a bit ridiculous it is a good way to alienate allot of donors from boinc

BOINC credits don't really mean anything. It's just a statistic. Different projects give out different amounts of credit for tasks completed. You don't really get anything from it; if anything it's just a way of saying how much work you've done.

They may not have a value to you but to many they do, I am not ashamed to look at the stats an see how I am doing compared to every body else out there, and I do take pride in seeing the assigned credit value of my contribution to the Boinc projects. All of my contribution is out of pocket there are no company funds or anything like that. According to the amount of credit I receive for my contribution my contribution of 400 cores and 4 GPU's to the Boinc project has become pretty minuscule compared to 1 ASIC miner.

I personally feel that my equipment is no longer wanted or needed in the Boinc projects and has been belittled by the amount of credit Bitcoin Utopia receives and to add to the rejection the ASIC miners are doing nothing as far as far as science goes, which is irrelevant but will affect my decision in the long run. I will not choose to run it, it is not my cup of tea, so I will have a choice to make somewhere down the line.

#1 either shut the rigs down
#2 move on to another project besides Boinc
#3 accept the credit disparity within the Boinc organization.

I will tell you now that the 3rd choice does not give me a warm and fuzzy feeling so that leaves #1 and #2

Anyway the statement that Boinc credits do not really mean anything may apply to you to to many of us out there in Bionc Land they do mean something and right at the moment by looking at the stats they mean the Boinc projects do not need any CPU's but they really need some ASIC miners.

In my opinion either the credit should be removed from Boinc stats and put in its own separate category outside of Boinc stats.

I personally feel that my equipment is no longer wanted or needed in the Boinc projects

I agree completely. GPUs have made my CPU-based contribution virtually worthless for a lot of projects. They are too old to accept decent GPUs, so now when they die I don't bother to repair them, I just throw them away. My single bottom-of-the-range GPU does as much work as most of the rest of my CPU-only machines put together. Depressing, isn't it? I'm just not interested enough to invest any money in a 'decent' machine (most of my stuff is $20 junk, my last couple of machines were free). On the bright side though if more work is being done by GPUs (even if not by me) that's got to be a Good Thing.

The whole credit thing is a red herring when in comes to cross-project comparisons. The best thing to do is just stop exporting stats. Just kill that function in the next BOINC server version, the stats sites will all be frozen in time, and everybody will be happy. Seriously, trust me on this.

I personally feel that my equipment is no longer wanted or needed in the Boinc projects

I agree completely. GPUs have made my CPU-based contribution virtually worthless for a lot of projects. They are too old to accept decent GPUs, so now when they die I don't bother to repair them, I just throw them away. My single bottom-of-the-range GPU does as much work as most of the rest of my CPU-only machines put together. Depressing, isn't it? I'm just not interested enough to invest any money in a 'decent' machine (most of my stuff is $20 junk, my last couple of machines were free). On the bright side though if more work is being done by GPUs (even if not by me) that's got to be a Good Thing.

YOU ARE MISSING THE POINT !!!!!
Credit is meant to be in proportion to the DRYSTONE/WHETSTONE rate of a device...
The actual indisputable truth is that the GPU:s are a LOT faster then CPU:s, and that is why they get more credit...
THIS IS SCIENTIFICALLY INDISPUTABLE!!!
TOOLS GET OBSOLETE!!!!

The problem with Bitcoin Utopia is that ELECTRICITY AND GEAR cost me and others a lot more than those morons gather... And yet, they get the most RESPECT AND RECOGNITION in terms of credits.

If any projects get the funds from them, I still do not think that
THEY CAN AFFORD 1000 HIGH END GPU WORKSTATIONS EVERY THIRD YEAR AND THE ELECTRICITY BILL THAT COMES WITH IT!!!

One word ---> LOL!

(I even DID IT IN CAPITAL LETTERS for you)

;)

One word ---> LOL!

(I even DID IT IN CAPITAL LETTERS for you)

;)

yeah, I know...

funny, how there is no place on the net where we can put our collective mental disorders behind us, and just get along on some good topic =)

EDIT:
when I was contributing the most, my power bill was like 35$/month higher then usual...
that is 420$/year just for crunching, and my rig was not even that elite...
For a 3 year period that is 800$ (gear) + 1260$ power = 2060$ for upper medium gear...

now, assume 500 crunchers: 500*2060=1.030.000$ in case of 24/7 contribution...

1 million dollars contributed by 500 crunchers over 3 years...

SOME ADDITIONAL STATS :D (not atypical numbers)
GPUGRID:
Short runs (2-3 hours on fastest card) 572 users
Long runs (8-12 hours on fastest card) 484 users

POEM:
POEM++ (cpu work units) 1,880 users
POEM++ OpenCL 315 users

so, assume 35$ more in power bill for 500 + 1800 users, and 3 years:
2300*35$*12*3=2.898.000$ power bill for medium hardware...

That is irrelevant as far as credit goes what is relevant is whether the projects under the Boinc projects need cpu's or not. Clearly the answer to that question is yes they do, because many of the projects are CPU only at this time. The problem lies in the Boinc projects lack of control or caring to control the awarding or granting of the credit within the Boinc Stats.

I see no problem with GPU's getting more credit than CPU's because they are doing more science what I do see a problem with is the ASCI miners getting so much credit, the ridiculous amount of credit makes anything else appear useless when compared to it and the only thing it is really doing is making a experimental form of currency, well I really do not know what it is doing nor does anybody else from what I can tell. The one thing I do know about Bitcoin mining is there has been plenty of scandal and corruption associated with it. But that is irrelevant also.

The question I have is does it belong in the Boinc Stats with it ridicules amount of credit granted. or should it be outside the Boinc Stats. I say it should be outside or lowered to reflect it's importance when compared to other projects and their needs.

I know my choices of where and what I run are guided by the credit system and my preferences, as I imagine some others do the same. I do not choose to run ASIC nor will I in my opinion the work they do is worthless to me. So from my prospective Boinc has become another ASIC Mining endeavour, at least it appears that way from the credit system.

most of what is said here is complete gibberish. the boinc credits are just a way to keep track of things. it would not matter much if one project had more points then another, except that it would show people using and contributeing to the project. regardless of who's team you joined you can mine and compute right alongside those projects and teams. it would not take away from anything.

asic vs cpu or gpu

also a bunch of people talking without much of anything being said. ALL data inside a computer is an integer...a number. no asics do any kind of floating point processing. unless they were made to. which many are not, and are way to far from being produced. In a manner people could easily use. beside a gfx card or general cpu, which does have a floating point processor.

All of them are the same, asics and gpus just throw data around in multiples slower then cpus currently do. if you notice, gpus and asic get tons more rejections and hw errors then cpu's do. But the randomly generate numbers are accepted in a more timely way. since most pools cater to what makes them money. hence the reasonings they put forward for excludeing cpu's. which is nonesense.

I for one wish my boinc credits and project credits earned me money. Then i could buy better hardware to both make MORE money and contribute more. all of this discrediting nonesense is just hurting yourselfs.

please let me know when gridcoin-qt is available again. Or of any projects that pay out cryptocoin of any sort. based on boic credits...

another thing is each project has their own point system outside of each other...so again it takes nothing from anyone or anything.

ALL hardware is needed. most projects are cpu only.

Its really sad when good things are pushed into the shadows, simply because people don't bother to understand before they degrade something they don't know.

The "contribution" of CPU vs GPU vs ASIC is also relative, since not all projects are able to make use of those higher-speed devices and run on CPU only. For some, this is because they don't have the development know-how, funding or time; for others it's because the computation being done cannot run or receives no boost for technical reasons. Even though they are very fast for some things, GPUs are not as flexible as a CPU--nor ASIC/FPGA, by definition.

I do keep track of my stats, but not for any competitive reasons--if I wanted to be a "competitive" BOINC cruncher, I would have to increase my income level significantly. You're simply not going to make any headroom against people who manage data centers, corporate IT pools or who have the funds to construct dedicated boxes chock full of GPUs and other high-end hardware. But, it is fun to look at how much I'm doing, even when just comparing my output from years past or to the few other (family) members of my team. The stats can be fun, but if that's the only reason you are computing for BOINC projects, then you're probably wasting your time and money. Seriously; thanks for your contribution, but the "green" thing about BOINC is that it can take advantage of unused resources on existing systems, reducing the need for assembling dedicated computing farms.

One thing that I've always appreciated about BOINC, is that even though a fancy top-of-the-line computer is going to outpace older ones, everyone's contribution is important. It's a collective, cooperative effort. Even if you just have one computer that's only on when you are actively using it, you are making a worthwhile contribution. You shouldn't need some arbitrary "currency" to prove that.

In the interest of disclosure, I do have Bitcoin Utopia attached (2 GPUs and miscellaneous CPUs) to support MilkyWay@home only. If the community decides that their stats should be hidden, so be it. Say what you will about the project, but they are directly supporting another BOINC project with funding.

---

My Detailed BOINC Stats

BOINC is, by design, a distributed loosely coupled network. This applies not only to the hosts that do the crunching but also to the projects that run the server software. Each project is a project on it's own. There is no authority that must or can authorize new projects. The success of a project is only derived from the number of users it can gather to crunch for it. So if someone is not interested in a specific kind of "science" don't attach to a project that does this.

The DRTG project is also highly disputed because it is no "science" and the output can be used to hack passwords. In my opinion the same applies to Bitcoin Utopia. If you don't like it don't crunch it.

Regarding the Credits issue with BU: There is currently a new design under discussion that should fix the problems that came with the advent of BU.
I would therefore like to point your attention to this page: CreditGeneralized in the developer wiki. If you have further comments on this new design please do so. You can do this here in this forum or via email.