Boinc's Death Knell?

Message boards : Politics : Boinc's Death Knell?
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 8 · Next

AuthorMessage
Sirius B Project Donor
Volunteer tester
Avatar

Send message
Joined: 26 Dec 00
Posts: 24870
Credit: 3,081,182
RAC: 7
Ireland
Message 1128557 - Posted: 17 Jul 2011, 1:10:17 UTC - in response to Message 1128398.  

Questions on MW policies or procedures should be asked on the MW forums.

Thank you very much
Blurf
MW Forum Mod/Project Admin
We Know SO why emphasis it?

This is one of the reasons why the bells are tolling....

..It tolls for Boinc!
ID: 1128557 · Report as offensive
Profile Sarge
Volunteer tester

Send message
Joined: 25 Aug 99
Posts: 12273
Credit: 8,569,109
RAC: 79
United States
Message 1128560 - Posted: 17 Jul 2011, 1:12:53 UTC - in response to Message 1128353.  

Why credit is screwed anyway.

It is based on a FLOP count, if that is counted accurately, and it is not. But it completely ignores the true amount of work done for a project. It ignores the cost of electricity and the cost of the network connectivity and the cost of the computing box. If you want a fair credit comparison then those factors must be included. It also doesn't reward the right people. If someone runs boxes at work on his account (with permission) the company is supplying the electricity, network connection and cost of the computers, but they are not getting the credit. I wonder is the IRS would want all that countable cash included on the W-2 form for the person getting the credit?


The Dean is aware I run BOINC on the two desktops in my office and that it is running when I bring my laptop to the office as well. I told him that if I was told to stop running it, I would comply, to save the university money for electricity. He said we probably weren't at that point yet, so just be sure I made wise choices for the screensaver settings, which I believe I already have.
ID: 1128560 · Report as offensive
Profile Gary Charpentier Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 25 Dec 00
Posts: 30593
Credit: 53,134,872
RAC: 32
United States
Message 1128569 - Posted: 17 Jul 2011, 1:51:57 UTC

Why credit can never accurately measure participation.

Some projects are strictly integer math. Some projects are strictly floating point math. Some projects are double precision floating point.

How much credit for an add instruction (8, 16, 32, or 64 bits) in each case?
How much credit for a subtract instruction in each case?
How much credit for a multiplication instruction in each case?
How much credit for a division instruction in each case?

What about a vector instruction?

What about a FFT instruction?

If you know anything about different CPU's you will understand that there is no way to equate these. If you can't equate them, then you can't add them into some equitable credit number. The best you can do is arbitrary. If the credit is already arbitrary, then why get upset about it no matter what any project does or offers?

I think that is the problem in a nut shell. There isn't a real equitable way to compare across projects how much work is donated.

ID: 1128569 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128593 - Posted: 17 Jul 2011, 3:33:14 UTC - in response to Message 1128538.  

Agreed -- I *can* see a roll for centralized 'arbitration' of severely out of balance credit schemes. I just don't know that CreditNew as (from the individual projects vantage point) as an externally imposed one size fits all plan is all that workable.

At a guess, it *may* be workable for 'standard' CPU projects -- though even there it will encounter significant issues regarding application design (which it doesn't seem to address), as well as both end user and project environmental factors. That being said it may be useful in that subset of the environment.

Frankly, I'm not at all sure the BOINC developers have that much hands on in the AMD/GPU side of things to make decent judgement calls there anyway.

I think the mess with Aqua may have been in part due to the special way its application works -- though it IS a CPU application, it is a multithreaded multi core application.


This phrase "have better things to do with their time" must not translate into "it's OK for the credit to be unfair and manipulated by some projects/individuals to their own advantage".


ID: 1128593 · Report as offensive
OzzFan Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Apr 02
Posts: 15691
Credit: 84,761,841
RAC: 28
United States
Message 1128598 - Posted: 17 Jul 2011, 4:26:08 UTC - in response to Message 1128555.  

I will step forward and state that I am someone who had his BOINC account merged with an older account from the S@H classic days. This was done by an admin. Eric, perhaps? No crocodile tears were involved.


Sorry, but the ability to "merge" a SETI Classic account to a BOINC account is available to everyone. If you were unaware or unable to make it happen and an Admin stepped in for you, they were only manually doing what is already available. It is important to note that this is not the same as merging two BOINC accounts, which is what is being discussed here.

When I opened my classic account, the e-mail address I used was a .edu account. Once someone leaves an educational institution, the e-mail account only remains open for something like 6 months to a year, depending on the institution and what you did there. (Student? Faculty?)


There's your mistake. You thought that the email needed to be a currently active one to join a Classic account to a BOINC account. In fact, the old email does not need to be active to "merge" a Classic account with a BOINC account.

Hopefully the discussion about "bad apples" is not including this being done for my account.


No, your situation is not what is being discussed as an issue.
ID: 1128598 · Report as offensive
OzzFan Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Apr 02
Posts: 15691
Credit: 84,761,841
RAC: 28
United States
Message 1128600 - Posted: 17 Jul 2011, 4:31:42 UTC - in response to Message 1128569.  

If you know anything about different CPU's you will understand that there is no way to equate these. If you can't equate them, then you can't add them into some equitable credit number. The best you can do is arbitrary. If the credit is already arbitrary, then why get upset about it no matter what any project does or offers?


For the same reason why humans try to make organization out of the natural chaos of the universe. We all have this sense of fairness that we want to achieve, and there's nothing wrong with coming as close to fair and non-arbitrary as we can.

I think that is the problem in a nut shell. There isn't a real equitable way to compare across projects how much work is donated.


The problem is that there's an expectation by some that it must be a perfect solution. We all know that perfection is out of the reach of anything we humans do, so we need to set the bar a little lower and actually make it obtainable as a goal in the interest of fairness.
ID: 1128600 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128608 - Posted: 17 Jul 2011, 4:47:50 UTC - in response to Message 1128600.  

For me I don't have an expectation of a perfect solution -- perhaps the folks powering CreditNew think their approach can be if not a perfect solution then a very good across the entire range of projects, applications, and processing devices solution. That would explain the push to fit all projects into that solution.

I think that CreditNew -- *an an option* makes a fair amount of sense. It might make more sense once it is actually ready for anything near prime time -- at least for 'straightforward' CPU BOINC projects. I think that increasingly pushing a 'not ready for prime time' scheme as the only option with new server process updates is likely to do CreditNew more harm than good - by seriously encouraging projects (particularly but not only GPU projects) to either pass on code updates (perhaps a bit of baby with bath water there), or go rogue by hacking the code updates to 'fix' the Force CreditNew code.

It also (as one can see from some of the discussion here) runs the risk of antagonizing at least some of the longstanding and active BOINC users who might be significantly more central control averse that BOINC central would seem to understand.





The problem is that there's an expectation by some that it must be a perfect solution. We all know that perfection is out of the reach of anything we humans do, so we need to set the bar a little lower and actually make it obtainable as a goal in the interest of fairness.


ID: 1128608 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128611 - Posted: 17 Jul 2011, 4:59:10 UTC - in response to Message 1128555.  

Ah -- OK -- I thought you were saying your classic 'credits' were merged into your account -- they are listed separately -- as our mine -- I was a busy camper even back then so I have 168,000 work units and 905,547 hours. I wonder what that would be worth in CreditNew credits <smile>


I will step forward and state that I am someone who had his BOINC account merged with an older account from the S@H classic days. This was done by an admin. Eric, perhaps? No crocodile tears were involved.
When I opened my classic account, the e-mail address I used was a .edu account. Once someone leaves an educational institution, the e-mail account only remains open for something like 6 months to a year, depending on the institution and what you did there. (Student? Faculty?)
If you look at my account, if it shows you, you will see "SETI@home classic workunits 2,329
SETI@home classic CPU time 81,529 hours."
This did not show up until late 2006 or some time in 2007. I mentioned it to someone, and apparently that someone passed on word to an admin (Eric?) and not long after that, the information showed up, accounts merged.
I may have had to state what the earlier e-mail address had been. I'm not sure if I was asked that.
Hopefully the discussion about "bad apples" is not including this being done for my account.


ID: 1128611 · Report as offensive
Profile GalaxyIce
Volunteer tester
Avatar

Send message
Joined: 13 May 06
Posts: 8927
Credit: 1,361,057
RAC: 0
United Kingdom
Message 1128623 - Posted: 17 Jul 2011, 6:42:49 UTC - in response to Message 1128600.  


The problem is that there's an expectation by some that it must be a perfect solution. We all know that perfection is out of the reach of anything we humans do, so we need to set the bar a little lower and actually make it obtainable as a goal in the interest of fairness.

Very well said and I agree. If it's not perfect due to the difficulty in making it perfect, at least we can try and make it fair and outlaw the practices by projects and individuals which are contrary to fairness.


flaming balloons
ID: 1128623 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128638 - Posted: 17 Jul 2011, 7:05:38 UTC - in response to Message 1128623.  
Last modified: 17 Jul 2011, 7:07:56 UTC

Or at reasonably equitable -- 'make it fair' depends on definition.

If you make 'fairness' into an absolute (and I may well be wrong but I hear that in your tone), then you have the potential of 'extremism in the service of fairness' -- which has a bit of the oxymoron to it.

But *outlaw* -- well I think that is a bit over the top -- *we* (note the collective instead of the singular) are not Congress. Then again, they make few laws these days....



Very well said and I agree. If it's not perfect due to the difficulty in making it perfect, at least we can try and make it fair and outlaw the practices by projects and individuals which are contrary to fairness.

ID: 1128638 · Report as offensive
Profile GalaxyIce
Volunteer tester
Avatar

Send message
Joined: 13 May 06
Posts: 8927
Credit: 1,361,057
RAC: 0
United Kingdom
Message 1128643 - Posted: 17 Jul 2011, 7:23:20 UTC - in response to Message 1128638.  

I would happily settle for equitable. As for outlaw - whatever term that says an unequitable practice of dishing out credits will be either not be possible (eg by any mechanism placed into BOINC) or reversed by compulsion if not voluntarilly by the project/individual when it comes to light. That would be fair enough I think.

Or at reasonably equitable -- 'make it fair' depends on definition.

If you make 'fairness' into an absolute (and I may well be wrong but I hear that in your tone), then you have the potential of 'extremism in the service of fairness' -- which has a bit of the oxymoron to it.

But *outlaw* -- well I think that is a bit over the top -- *we* (note the collective instead of the singular) are not Congress. Then again, they make few laws these days....



Very well said and I agree. If it's not perfect due to the difficulty in making it perfect, at least we can try and make it fair and outlaw the practices by projects and individuals which are contrary to fairness.



flaming balloons
ID: 1128643 · Report as offensive
Sirius B Project Donor
Volunteer tester
Avatar

Send message
Joined: 26 Dec 00
Posts: 24870
Credit: 3,081,182
RAC: 7
Ireland
Message 1128761 - Posted: 17 Jul 2011, 15:37:53 UTC - in response to Message 1128608.  

[quote]snip

I think that CreditNew -- *an an option* makes a fair amount of sense. It might make more sense once it is actually ready for anything near prime time -- at least for 'straightforward' CPU BOINC projects. I think that increasingly pushing a 'not ready for prime time' scheme as the only option with new server process updates is likely to do CreditNew more harm than good - by seriously encouraging projects (particularly but not only GPU projects) to either pass on code updates (perhaps a bit of baby with bath water there), or go rogue by hacking the code updates to 'fix' the Force CreditNew code.

snip
[quote]

Very well stated. This is why I feel that Boinc Alpha should be restarted. There are many of us out there who will assist in testing, regardless of credit or the lack of it.

Get Boinc right & I don't think that there will be many more debates like this current one.

Dr A has done great on the original concept, so I personally feel that he has to step up to the plate & ensure it continues to work as he envisaged it.

Currently, it is not!

Boinc only remains on my workstation while primaboinca finishes its tasks. once that is completed, Boinc will no longer reside on my network & I'll only return when irritating coding problems like the current boinc screensaver diagnostic error no longer appear & boinc does what it's supposed to, rather than running as a beta test.
ID: 1128761 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1128767 - Posted: 17 Jul 2011, 15:43:19 UTC - in response to Message 1128761.  
Last modified: 17 Jul 2011, 15:43:40 UTC

Boinc only remains on my workstation while primaboinca finishes its tasks. once that is completed, Boinc will no longer reside on my network & I'll only return when irritating coding problems like the current boinc screensaver diagnostic error no longer appear & boinc does what it's supposed to, rather than running as a beta test.

Then I guess you'll never run Boinc again, as it will ALWAYS be a work in progress, and always have a few bugs.
By and large, Boinc DOES do what it is supposed to do.
Millions upon millions of WUs processed and returned to the various projects under the Boinc umbrella testify to that.

If you wish to contribute to distributed computing, accept Boinc with it's foibles.

If not, that is your choice, my friend.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1128767 · Report as offensive
Sirius B Project Donor
Volunteer tester
Avatar

Send message
Joined: 26 Dec 00
Posts: 24870
Credit: 3,081,182
RAC: 7
Ireland
Message 1128783 - Posted: 17 Jul 2011, 15:59:07 UTC
Last modified: 17 Jul 2011, 15:59:26 UTC

Fair point Mark & totally understand & agree. However, if boinc is to continue as a test bed, please explain why Boinc Alpha came into being?

Wasn't the intention to use that as a test bed for boinc, while boinc main continued to do the work it was designed to do?

Why stop it? This creditnew coding is not working - it's interfering with projects which imo is defeating the originial purpose of boinc.

Most companies & individuals that test normally use a "test lab" for any/all new software/hardware before it ever enters production.

All that I can see from the boards are those for & those against "credit". any mention of problems & decent debates evolve into a credit war.

The orginial point of this thread was regarding the boinc coding & it's failures.

AND, so far, I've seen nothing that homes in on those points!
ID: 1128783 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1128790 - Posted: 17 Jul 2011, 16:05:13 UTC - in response to Message 1128783.  
Last modified: 17 Jul 2011, 16:05:51 UTC

Fair point Mark & totally understand & agree. However, if boinc is to continue as a test bed, please explain why Boinc Alpha came into being?

Wasn't the intention to use that as a test bed for boinc, while boinc main continued to do the work it was designed to do?

Why stop it? This creditnew coding is not working - it's interfering with projects which imo is defeating the originial purpose of boinc.

Most companies & individuals that test normally use a "test lab" for any/all new software/hardware before it ever enters production.

All that I can see from the boards are those for & those against "credit". any mention of problems & decent debates evolve into a credit war.

The orginial point of this thread was regarding the boinc coding & it's failures.

AND, so far, I've seen nothing that homes in on those points!

Well, Sirius...
It's your thread, so if the direction of the discussion does not suit you, post about it as you just did. If that does not redirect things, the red-x is a last resort to ask the mods to intervene on your behalf. Those are your privileges as thread owner.

As to Boinc Alpha....
I rather suspect there are far too few people participating for the devs to get enough testing across a wide enough range of variables for it to give a valid enough litmus test.
So I suspect that's why Seti becomes Boinc Alpha much of the time.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1128790 · Report as offensive
Sirius B Project Donor
Volunteer tester
Avatar

Send message
Joined: 26 Dec 00
Posts: 24870
Credit: 3,081,182
RAC: 7
Ireland
Message 1128802 - Posted: 17 Jul 2011, 16:14:51 UTC

Thanks Mark, another 2 good statements.

1: I prefer open communications. Even a bad post can often contain a snippet of a solution to a particular problem (& there also make good reading...lol).

2: I wasn't aware of that fact. However, BarryAZ's point comes to mind regarding primetime.

So as the devs are volunteers themselves, WHY NOT just run the changes on their own machines - surely they would be able to see the problems that crop up?

Lunatics do it so I can't see why Boinc Devs/Admin can't do it.
ID: 1128802 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128840 - Posted: 17 Jul 2011, 17:29:11 UTC - in response to Message 1128643.  

OK -- reasonably equitable sounds right.

Regarding outlaw -- over time, we've seen the effects of various forms of pressure which seem to have done a reasonably effective job regarding credit systems.

I think the issue there is that some of the hardware (particularly GPU hardware) can be utilized in ways that may well *legitimately* produce large credit numbers. I suspect that some of the thrust behind CreditNew is to 'control' the GPU project scores as they are seen by some folks at BOINC central as 'disturbing the force'. The problem here is that it appears that most of the force behind this comes from folks who don't have that much experience with designing applications for GPU use, especially AMD GPU use.

As to voluntarily reducing credit numbers generated -- in part define voluntary, and in part the question of application optimization. When a project starts at a certain 'credit level' and then over time optimizes code, should the credits for the same work unit be reduced because of the greater efficiency? I think that presents a problem as well as it, in a sense, eliminates the incentive to improving application code. That surely would be an unintended negative consequence.


I would happily settle for equitable. As for outlaw - whatever term that says an unequitable practice of dishing out credits will be either not be possible (eg by any mechanism placed into BOINC) or reversed by compulsion if not voluntarilly by the project/individual when it comes to light. That would be fair enough I think.




ID: 1128840 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1128841 - Posted: 17 Jul 2011, 17:35:53 UTC - in response to Message 1128802.  
Last modified: 17 Jul 2011, 17:46:26 UTC


So as the devs are volunteers themselves, WHY NOT just run the changes on their own machines - surely they would be able to see the problems that crop up?

Lunatics do it so I can't see why Boinc Devs/Admin can't do it.

They very well may, and could catch major bugs. But how many combinations of OS, CPUs, GPUs, RAM, etc., etc., etc. could they possibly test with such a small sampling? I am sure some in house testing does get done...then on to the limited pool of Alpha testers...and then it goes live on Seti.
I suspect that some bugs don't show themselves until you have thousands of computers involved in the 'test'.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1128841 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1128845 - Posted: 17 Jul 2011, 17:59:31 UTC

And...LOL...before somebody starts calling me a Boinc fanboi...

There are things I don't like about Boinc either.
Like the various combinations of retry backoffs, project backoffs, suppressing work requests just because you have too many pending uploads or downloads in retry status. And the backoff times in current versions of Boinc are even longer than in the version I run. They keep me from getting work when comms with the servers are, shall we say, less than optimum. But I guess they do help keep the servers from melting down totally.

If I were to compile my own version of Boinc......
Well, let's just say it would be one server hammering SOB.

LOL.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1128845 · Report as offensive
BarryAZ

Send message
Joined: 1 Apr 01
Posts: 2580
Credit: 16,982,517
RAC: 0
United States
Message 1128846 - Posted: 17 Jul 2011, 18:10:06 UTC - in response to Message 1128845.  

Which I suppose could be used as a rational NOT to allocate sparse resources on a NewCredit scheme which adds all those variables (and others) to the 'please test before release queue'.




There are things I don't like about Boinc either.
Like the various combinations of retry backoffs, project backoffs, suppressing work requests just because you have too many pending uploads or downloads in retry status. And the backoff times in current versions of Boinc are even longer than in the version I run. They keep me from getting work when comms with the servers are, shall we say, less than optimum. But I guess they do help keep the servers from melting down totally.



ID: 1128846 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 8 · Next

Message boards : Politics : Boinc's Death Knell?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.