I would be interested if you could make an ordered list of what, in your opinion, are the most important problems with BOINC, SETI, and distributed computing projects. The BOINC community will be meeting in France in a couple of weeks and this would be an interesting thing for me to bring to that discussion.
Cheers,
Bruce
High on my list would be :
* Abandon the ideal of cross project credit parity
* put an RSS reader into BOINC manager so that crunchers get project news delivered right to the GUI (maybe as a tickertape like display...just an idea)
I really think someone should write up a memo "Cross project credit parity considered harmful" along the lines of Mike's suggestion at the end of this message :
Quote:
[personal view]
Actually I've seriously & often wondered if cross-project credits are usefully meaningful. I think it really compares apples, oranges and deep-sea squids. In the long run it may be better to define credit currencies - E@H credits, SETI credits, Cosmology credits etc. What/how/when contributors think of cross project parity is then their own pleasure, you'd be free to define whatever exchange rates you like. After all, not all 'fly-buy' schemes are equivalent for much the same reasons.
[/personal view]
Cheers, Mike.
BOINCland would be a more peaceful place without all the irritations caused by the credit adjustments that are indeed inevitable as long as cross-project parity is considered a goal.
On a more constructive note, I'd love to see a lightweight RSS reader being integrated into the BOINC_manager GUI to stream the project news from the project homepages to the users. Would make communication of important messages much easier, I think. Just providing a button that will open the Homepage is not enough to keep the users involved and up to date IMHO.
[personal view]
Actually I've seriously & often wondered if cross-project credits are usefully meaningful. I think it really compares apples, oranges and deep-sea squids. In the long run it may be better to define credit currencies - E@H credits, SETI credits, Cosmology credits etc. What/how/when contributors think of cross project parity is then their own pleasure, you'd be free to define whatever exchange rates you like. After all, not all 'fly-buy' schemes are equivalent for much the same reasons.
[/personal view]
I stand by that. It leaves contributors utterly free to assign their own personal valuations with respect to different projects, without impost on any other user what-so-ever. It would eliminate at a stroke a vast source of disquiet, and allow all involved parties to sharpen focus on other more productive aspects of distributed computing, by terminating an unachievable task - that of officially validating fundamentally disparate value judgments.
So make that #1 in a list of one(1) items from me .... :-)
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
I agree with these views concerning cross-project credit parity. As far as I'm concerned it simply isn't achievable and the attempts to achieve it create a huge amount of angst.
I get fun out of tweaking things to make measurable improvements. Switch machines to Linux - I see a measurable improvement. Install the latest optimised app - have fun measuring what the improvement is. Replace some PIII coppermine 1Gig boxes with tualatin celeron 1300s - see the big improvement. Then along comes a cross-project credit parity adjustment and everything gets destroyed for a period of several weeks to months or more until things settle down again. The continuity of what is going on is lost and it leaves a very jarring and unsatisfactory feeling.
I must say that the second half of R3 when the optimised apps started to bite and credit was left alone was a very enjoyable time indeed. I had a lot of fun pushing my RAC to the highest it had ever been. I couldn't care less about what the rates of credit on other projects are. Each project is quite a separate entity and most people choose projects based on their perception of the quality and worth of the project and its personalities and performance rather than on any credit/hr metric.
The key point is stability of the credit granting mechanism within a given project as far as I'm concerned. What other projects might be offering isn't even on my radar. If it were, I wouldn't be trying to crunch for LHC on the rare occasions it has work :-).
I agree with these views concerning cross-project credit parity. As far as I'm concerned it simply isn't achievable and the attempts to achieve it create a huge amount of angst.
I agree also. I look at the credit I get within a project, and don't really think about how it equates to other projects I may also be working on.
Realistically, if a project were to give way too low or way too high compared with other projects, there may be an outcry. But that can be a self-regulating mechanism by the crunchers themselves, not a "dictate from above".
I guess a "free market" idea is what I think would work well! Let's give it a shot and see if it works (and I think it will!).
If you remove the idea of cross project credit parity, therefore surely the next step is to remove credits and all other measuring systems totally. And when I say totally I mean just that no accounting for work done even within the project. No counting of units done, because 10 S5R4's (oranges) do not equal 10 S5R3's (apples). No totaling of time done etc. etc.
But remember there are a lot of competitive people out there, and I'm sure that will bring just as much angst and questions. People still ask "How can I find out how many tasks have I completed?"
One team constantly advertises it is doing 22.4 TerraFlops, 22.3cr/sec.
And there is a BOINCTeams.com website.
And looking at the credits on Seti, then it looks like Eric's method is working, as far as I can see on the multibeam, common angle range, tasks the credits have reduced ~ 12% so far. And on Astropulse units the credits are increasing, but as they can take >100hrs it is too soon to give a percentage figure for the change.
My main emphasis on what I said was intended to be more to SETI-specific problems, considering he (David) is still listed as Director. I do not have handy a list of specific "problems" with the version of BOINC that I use on my computer, but I have had problems with trying to update to newer versions, that's why I stick with 5.8.16. Also, various problems have come into play with forum changes that seem to be slid in with other updates. Something that comes to mind whas when avatars were brought into forum "Author" fields when viewing the list of messages, that and the whole "Friends" concept causing significantly more bandwidth if you're someone like Misfit with a long list of friends...
Anyway, since I stay back at 5.8.16 and I don't use the social networking functions, I don't have a specific "BOINC" problem list, but I can mill around looking at Trac or perhaps asking some of the more "seasoned" people as far as software and back-end stuff. I'll see what I can come up with.
Oh, right off the top, I'd want to be sure that the long-standing NTLM proxy problems with the communications section of the client has been fixed.
As to what others have said about credit "parity", my thoughts on this are that David appears to have it as a "pet peeve" type of item. The data does not show that there are these large numbers of people that move from one project to the next based on just their credit awarded. Yes, there are some people who do exactly that, but it just isn't as big of an issue as what it is made out to be. I think that the "calibrating" BOINC clients that messed with the benchmark values is what got David fixated on the issue. There was a lot of negativity during that timeframe, especially in regards to allegations of whether someone was 'cheating' or that they were 'cherry-picking', etc, etc, etc... I think it may be time to take an objective look at the situation and see if the problem really deserves the effort being put into it. More than that, it deserves an objective look to see if the efforts are really going in the right direction, or if they will just be a temporary placebo...
If you remove the idea of cross project credit parity, therefore surely the next step is to remove credits and all other measuring systems totally. And when I say totally I mean just that no accounting for work done even within the project. No counting of units done, because 10 S5R4's (oranges) do not equal 10 S5R3's (apples). No totaling of time done etc. etc.
Take some time and work through Alinator's questions in the 64000 cobblestone questions thread and then remember both of my examples (widgets and identical numbers of tasks performed during different credit schema eras).
The point that a lot of you pro-CPP folks don't seem to understand is that the actions of all current CPP proposals are deflationary over time. All that you all seem to see is that it is "anti-inflationary". You simply will not see the deflationary aspects.
Example:
If my computer is the same today as it was a year ago and the application is the same today as it was a year ago, but the "median" computer has gone up in performance, my system performing the exact same work in the exact same amount of time would end up being awarded less due to the recalibration caused by bringing the "power" aspect (what Alinator is referencing) of the newer systems into the equation. Clearly if my system is below the median, I should get below-median compensation, right? The further away from the median I become, the less the compensation becomes, right?
In the BOINC Survey, when asked for reasons to join a project,
the answer "because I get more credits there than in other projects" (translated from German) received the lowest numbers of votes among all alternatives. The answer "Because the Research is important and beneficial to the public" got more than 20 times as many votes.
To be fair, the "cross project credit-aware" users will probably have a more than proportional share in the overall project performance, but still I wonder whether this cross-project credit parity thing is too big an effort for a very, very small (but outspoken) minority of the user base, confusing/annoying quite a lot of people at the same time.
In the BOINC Survey, when asked for reasons to join a project,
the answer "because I get more credits there than in other projects" (translated from German) received the lowest numbers of votes among all alternatives. The answer "Because the Research is important and beneficial to the public" got more than 20 times as many votes.
CU
Bikeman
Trying to think this through from all perspectives, this issue did not seem to have been an issue before Mr. Anderson brought it up--is that correct? Things seemed to have been working fine, with no major complaints as far as I know.
So returning to that more "open" system where each project acts as their own "policeman" would seem to be OK, afaik.
RE: France? I thought it
)
That's a different thing. Bruce is talking about the 4th Pan-Galactic BOINC Workshop.
RE: I would be interested
)
High on my list would be :
* Abandon the ideal of cross project credit parity
* put an RSS reader into BOINC manager so that crunchers get project news delivered right to the GUI (maybe as a tickertape like display...just an idea)
I really think someone should write up a memo "Cross project credit parity considered harmful" along the lines of Mike's suggestion at the end of this message :
BOINCland would be a more peaceful place without all the irritations caused by the credit adjustments that are indeed inevitable as long as cross-project parity is considered a goal.
On a more constructive note, I'd love to see a lightweight RSS reader being integrated into the BOINC_manager GUI to stream the project news from the project homepages to the users. Would make communication of important messages much easier, I think. Just providing a button that will open the Homepage is not enough to keep the users involved and up to date IMHO.
CU
Bikeman
RE: [personal
)
I stand by that. It leaves contributors utterly free to assign their own personal valuations with respect to different projects, without impost on any other user what-so-ever. It would eliminate at a stroke a vast source of disquiet, and allow all involved parties to sharpen focus on other more productive aspects of distributed computing, by terminating an unachievable task - that of officially validating fundamentally disparate value judgments.
So make that #1 in a list of one(1) items from me .... :-)
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
I agree with these views
)
I agree with these views concerning cross-project credit parity. As far as I'm concerned it simply isn't achievable and the attempts to achieve it create a huge amount of angst.
I get fun out of tweaking things to make measurable improvements. Switch machines to Linux - I see a measurable improvement. Install the latest optimised app - have fun measuring what the improvement is. Replace some PIII coppermine 1Gig boxes with tualatin celeron 1300s - see the big improvement. Then along comes a cross-project credit parity adjustment and everything gets destroyed for a period of several weeks to months or more until things settle down again. The continuity of what is going on is lost and it leaves a very jarring and unsatisfactory feeling.
I must say that the second half of R3 when the optimised apps started to bite and credit was left alone was a very enjoyable time indeed. I had a lot of fun pushing my RAC to the highest it had ever been. I couldn't care less about what the rates of credit on other projects are. Each project is quite a separate entity and most people choose projects based on their perception of the quality and worth of the project and its personalities and performance rather than on any credit/hr metric.
The key point is stability of the credit granting mechanism within a given project as far as I'm concerned. What other projects might be offering isn't even on my radar. If it were, I wouldn't be trying to crunch for LHC on the rare occasions it has work :-).
Cheers,
Gary.
RE: I agree with these
)
I agree also. I look at the credit I get within a project, and don't really think about how it equates to other projects I may also be working on.
Realistically, if a project were to give way too low or way too high compared with other projects, there may be an outcry. But that can be a self-regulating mechanism by the crunchers themselves, not a "dictate from above".
I guess a "free market" idea is what I think would work well! Let's give it a shot and see if it works (and I think it will!).
If you remove the idea of
)
If you remove the idea of cross project credit parity, therefore surely the next step is to remove credits and all other measuring systems totally. And when I say totally I mean just that no accounting for work done even within the project. No counting of units done, because 10 S5R4's (oranges) do not equal 10 S5R3's (apples). No totaling of time done etc. etc.
But remember there are a lot of competitive people out there, and I'm sure that will bring just as much angst and questions. People still ask "How can I find out how many tasks have I completed?"
One team constantly advertises it is doing 22.4 TerraFlops, 22.3cr/sec.
And there is a BOINCTeams.com website.
And looking at the credits on Seti, then it looks like Eric's method is working, as far as I can see on the multibeam, common angle range, tasks the credits have reduced ~ 12% so far. And on Astropulse units the credits are increasing, but as they can take >100hrs it is too soon to give a percentage figure for the change.
RE: Bruce's request My
)
RE: Bruce's request
My main emphasis on what I said was intended to be more to SETI-specific problems, considering he (David) is still listed as Director. I do not have handy a list of specific "problems" with the version of BOINC that I use on my computer, but I have had problems with trying to update to newer versions, that's why I stick with 5.8.16. Also, various problems have come into play with forum changes that seem to be slid in with other updates. Something that comes to mind whas when avatars were brought into forum "Author" fields when viewing the list of messages, that and the whole "Friends" concept causing significantly more bandwidth if you're someone like Misfit with a long list of friends...
Anyway, since I stay back at 5.8.16 and I don't use the social networking functions, I don't have a specific "BOINC" problem list, but I can mill around looking at Trac or perhaps asking some of the more "seasoned" people as far as software and back-end stuff. I'll see what I can come up with.
Oh, right off the top, I'd want to be sure that the long-standing NTLM proxy problems with the communications section of the client has been fixed.
As to what others have said about credit "parity", my thoughts on this are that David appears to have it as a "pet peeve" type of item. The data does not show that there are these large numbers of people that move from one project to the next based on just their credit awarded. Yes, there are some people who do exactly that, but it just isn't as big of an issue as what it is made out to be. I think that the "calibrating" BOINC clients that messed with the benchmark values is what got David fixated on the issue. There was a lot of negativity during that timeframe, especially in regards to allegations of whether someone was 'cheating' or that they were 'cherry-picking', etc, etc, etc... I think it may be time to take an objective look at the situation and see if the problem really deserves the effort being put into it. More than that, it deserves an objective look to see if the efforts are really going in the right direction, or if they will just be a temporary placebo...
RE: If you remove the idea
)
Take some time and work through Alinator's questions in the 64000 cobblestone questions thread and then remember both of my examples (widgets and identical numbers of tasks performed during different credit schema eras).
The point that a lot of you pro-CPP folks don't seem to understand is that the actions of all current CPP proposals are deflationary over time. All that you all seem to see is that it is "anti-inflationary". You simply will not see the deflationary aspects.
Example:
If my computer is the same today as it was a year ago and the application is the same today as it was a year ago, but the "median" computer has gone up in performance, my system performing the exact same work in the exact same amount of time would end up being awarded less due to the recalibration caused by bringing the "power" aspect (what Alinator is referencing) of the newer systems into the equation. Clearly if my system is below the median, I should get below-median compensation, right? The further away from the median I become, the less the compensation becomes, right?
In the BOINC Survey, when
)
In the BOINC Survey, when asked for reasons to join a project,
the answer "because I get more credits there than in other projects" (translated from German) received the lowest numbers of votes among all alternatives. The answer "Because the Research is important and beneficial to the public" got more than 20 times as many votes.
To be fair, the "cross project credit-aware" users will probably have a more than proportional share in the overall project performance, but still I wonder whether this cross-project credit parity thing is too big an effort for a very, very small (but outspoken) minority of the user base, confusing/annoying quite a lot of people at the same time.
CU
Bikeman
RE: In the BOINC Survey,
)
Trying to think this through from all perspectives, this issue did not seem to have been an issue before Mr. Anderson brought it up--is that correct? Things seemed to have been working fine, with no major complaints as far as I know.
So returning to that more "open" system where each project acts as their own "policeman" would seem to be OK, afaik.