FormulaBoinc Forum

Michael H.W. Weber wrote: ↑Wed Jan 10, 2024 7:52 pmIn case people read this who are interested in developing novel BOINC apps & projects, please contact us. We have plenty of cool ideas.

Hallo Michael,

it's nice to know that I'm not the only one with ideas and engagement in BOINC. I will definitely contact you later after we will finish our top priority project for now - Defeatrussia@home

I hope that we are still able to set the world to make science, not war.

Beste Grüße
Dmytro

After another three people from our team announced yesterday to no longer continue computing for the team, I think it might be of interest to point out another issue with the FB changes - which we debated in our forum sometime around June 2022: It is the change to a system in which tactical "bunkering" of tasks is not longer of advantage.
As a consequence, FB has become a pure "Materialschlacht" as we say in Germany, meaning that only the vast compute power counts, no more focused upload tactics and the like - making the competitions increasingly boring. And that, at least in our team, has resulted to more and more people quitting BOINC.
Maybe it would be a good idea to turn back to the old system?

Michael.

P.S.: I understand the current rules also were a measure to reduce server load.

I think tactical 'bunkering' makes a lot of fun with strategy variations only for events like BOINC Pentathlon. The idea of sprint is really great to run a challenge for a certain period. Today I've spent almost the whole day setting up bunkers for SRbase challenge for all machines which compromised my German class for today. I didn't like that. Especially when BOINC scheduler behaves unpredictably strange calculating the number of tasks to download.

Anyway, we need to set up some exciting and dynamic competition in the same time not so complex to run at participant's side.

Michael H.W. Weber wrote: ↑Thu Jan 11, 2024 3:42 pm After another three people from our team announced yesterday to no longer continue computing for the team, I think it might be of interest to point out another issue with the FB changes - which we debated in our forum sometime around June 2022: It is the change to a system in which tactical "bunkering" of tasks is not longer of advantage.

Hi Michael

The issue of bunkering (and specifically related to FB Sprints) is that some teams employed some very unfriendly tactics such as the following:

1) Changing their BOINC Manager cache so as to download excessive number of tasks, (sometimes more than the 10+10 days allowed in the BOINC Manager settings) well in advance of the start date and time, leaving few tasks available for people who did NOT bunker.

2) Leaving it until the last minute to UPLOAD completed tasks, resulting in increased server load and preventing other members uploading their completed tasks as well as this limiting the validators from actually awarding BOINC credits.

3) There was a huge number of tasks that were abandoned once the Sprint ended, which meant other people taking part in the Sprint (who were part of the same quorum for those tasks) would not have their tasks validated until these abandoned tasks were re-issued...so they received NO BOINC credits (and therefore no chance of earning FB points) for their efforts, until well after the Sprint ended.

So, this "gaming of the system" might have helped those who wanted to prove a point, but it actually did not help the projects themselves, so the projects lost out due to this.

Maybe in the future, for Formula One races, (on which FB is based) you would prefer some drivers to start earlier so that they would be racing on the circuit BEFORE the official start time of the actual race...they can then warm up their engines, get their tyres in the sweet spot and maybe complete a few laps that count towards the actual race distance? Is that fair?

An FB Sprint should and does have an official start and end time...so that keeps everything as fair as possible with no one gaining an advantage due to time differences or being able to download tasks in advance. As it is now, it is as fair as it needs to be...and if people cannot get an unfair advantage then more people will hopefully take part. And as for those cheats who want to buck the system?

Michael H.W. Weber wrote: ↑Thu Jan 11, 2024 3:42 pm As a consequence, FB has become a pure "Materialschlacht" as we say in Germany, meaning that only the vast compute power counts, no more focused upload tactics and the like - making the competitions increasingly boring. And that, at least in our team, has resulted to more and more people quitting BOINC.
Maybe it would be a good idea to turn back to the old system?

Michael.

P.S.: I understand the current rules also were a measure to reduce server load.

Yes, it is true that *all* challenges (not just on FB) are liable to attract members who have lots of resources at their disposal.

But on the other hand, isn't it the responsibility of the team founder, as well as any team "cheerleaders" to actually promote such Challenges within their respective teams and thereby generate lots of team member participation for each Challenge?

And at the end of the day, FB should be about having some fun and generating some competitive spirit to *assist the projects* to generate a lot of completed and valid tasks in a short period of time...the whole issue of getting points should be secondary to this...

Boring competitions are only boring when the same people turn up and take part....such as Bayern Munich, Real Madrid or Manchester City always winning everything each season. It's when a team like Leicester City come from nowhere and win the Premier League that other teams take notice and see what *they* could achieve, if only they tried that bit more !

Ultimately, FB is about helping the projects to achieve their results more quickly by team members focusing on one project for a short period of time...anything else achieved is just "gloss" and has no real meaning in actual life.

It's not like the 2023 Marathon winners suddenly become millionaires or have schools named after them or are awarded Nobel prizes...

regards
Tim

Hi again

Just another thing to point out:

Using BOINCstats you can see how many users have earned BOINC credits in the last month and the results are quite staggering.
There could be many duplicate accounts counted more than once in the above "registered" numbers figures, but the downshift in the number of people actually crunching, compared to say 10 years ago, (or even 5 years ago) is worrying.

And whether BOINC can withstand this level of apathy is also quite a problem, as without a significant pool of users (from which to draw resources from) there is no point projects adopting BOINC as a mechanism to generate the results they need.

Ideally of course, it would be great to generate more interest in BOINC...but given the increase in charges for energy around the world and the need for people to cut back on non-essentials, means that for now, BOINC is going to struggle to keep enough people interested, donating their resources and paying for the energy used....it is just a shame that such energy costs to run BOINC cannot be deducted from ones taxable income, similar to charity donations that are tax-deductable in some countries?

regards
Tim

Just three points:

1) I have no problem letting you have your opinion on the way BOINC competitions should be organized. However, you should not portray people who evaluate things differently as "unfair" or "cheaters" (and I believe you really didn't mean it that way).
On to a few aspects: Let's take Primegrid competitions: there, only tasks that have been downloaded from the start of the competition are counted. So premature task downloading is not a problem at all, it can be handled. Everything else you mentioned is ultimately related to the technical performance of the BOINC project servers and the internet connectivity of the clients and you have to be able to assess this - as part of your team's competition tactics. And that is not so simple.
You yourself mentioned the example of football clubs, where you favor the small, less well-heeled teams winning (I agree on that). In the BOINC world, these are also the smaller teams with less computing power. And if they win, or at least deliver a particularly good result, they have to be particularly clever because otherwise they would be crushed by the sheer computing power of the others.
With the current FB system, this is simply impossible. The winners are always those with the most computing power. And that, as you described for the football club situation, gets boring at some point.
And that was exactly the point the teammates wanted to make in this discussion - nothing more, nothing less.

2) To my opinion there is no point in highlighting that the projects do suffer when people use tactics during competition. And that is for the simple reason that we do compute for BOINC projects all year long. In fact, many projects have way less tasks than we can process. So why shouldn't we as the funders of all the BOINC projects also have a little fun during the few competition periods? And it is not that tasks that don't count in the competition are lost or are not validated. As long as they are submitted within the project's task deadline, everything is fine.

3) You suggest that the decline in BOINC participation is a result from the surge of energy pricing. Could be - at least in part. But when I asked my team mates, I received a different reply: (a) It is the lack of good and thematically diverse BOINC projects (which declined over the period where interest in BOINC declined as well). And (b) it is the lack of interestingly organized, challenging competitions that are not just summing up the sheer compute power (as FB now also does).
The latter is also the reason for why the BOINC Pentathlon is eagerly awaited by the BOINC community every year.

So, that's just a few thoughts on the topic.

Michael.

P.S.: As the discussion develops, I don't think FB will be changed. That's completely ok with me - it is ultimately your system. I just wanted to bring a few thoughts over to this board which we discussed on ours. And maybe someone else will add another competition system to provide a bit more of diversity in the field, which would be good I think.

Hi Michael

Thank you SO MUCH for your considered opinions and thoughts, which I warmly welcome - it is so nice to be able to discuss these sorts of topics openly and in an adult and straight forward way - after all, we all want to help the projects in what ever way we can, so it is best for everyone if we band together and do what we can to provide some resources for the projects.

re: My "cheaters" comment

When FB started the Sprints, which was before I became involved, it was, I think expected that FB members would only start crunching tasks as from the official start date and time - this I thought was a "gentlemans agreement" situation so everyone started at the same time...even though FB gave advance warning as to the name of the project. And it was this that allowed members to bunker tasks in advance of the official start time. As such, some members ignored the "gentlemans agreement" and hence they had a head start, which seems a bit unfair to my mind.

And to try and balance this out, this is why Sprint projects are now announced at the same time as the official start time AND also that FB points are now awarded on each day of the Sprint, depending on team placings, which then lessens the issue of overloading the project server for the last hour or two of the Sprint and so completed tasks have more chance of being validated within the Sprint window.

You mention about PrimeGrid and their Challenges...and this is very successful and is run very well. But they can set specific times for the start and end of their Challenge as they know the date and time when tasks are downloaded and so they can filter out (within their statistics) any tasks that were downloaded in advance and they can exclude them.

It is my understanding that FB has no way of doing this, as it only collects the user name, team name and BOINC credits awarded on each day and there is no data exported for when the tasks were first downloaded.

re: Size of teams

Yes, some teams have done very well to generate lots of member interest (in the FB Challenges) and some of those members do have large farms of hosts which can help during Sprints. There's not much one can do about this but it is still perhaps in competing teams interest to actually get more individuals to join in, even if only for a day or two, and then little teams could actually battle some of the bigger teams,

re: The future of FB and it's possible development.

I like to think that FB is NOT *MY* system and that I have been purely the curator for a while, and I have just been trying to keep "all the plates spinning and not crashing on the ground". With Sebastien seeming to want to take a far more "hands off" approach, there are simply two outcomes: it either shuts down OR someone who is interested in keeping it going, can take over some or all of the administration and future direction.

For the time being, I am happy to try and make the latter option the better choice...and for now, unless someone else has a desire to take over, then I will just try and keep it running and maybe with a few tweaks here and there to keep up the members interest.

But in the short term, my current plan is purely to keep it going as it is, and if other ideas are put forward then it is a question of how such ideas can be implemented and at what cost (in time and labour as well as any programming costs).

And if anyone else wants to make any suggestions then please feel free to do so...this forum is about listening to members, providing feedback and guidance with regards to the FB Marathon and Sprints as well as generating interest in BOINC projects.

with best wishes
Tim

My own personal opinion about the decline of boinc over the years is not due to precise technical reasons or only a matter of publicity, for me it parallels the hysterical evolution of our societies : it's a time of turning inward attitudes, excessive nationalisms, "exclusionism" and success of populism, everywhere. Mainly because perspectives are darks, global warming and rising totalitarian powers, not talking about generalize economical difficulties and energy cost increase (and I do think active cruncher reducing or stopping their boinc participation is really linked to this) for the majority of us (*) make it hard to look to the future with a smile.

From my point of view Boinc is a bit of the opposite of all that : selfless help with the idea of a bright future, credit race being only about amusement around it to keep a community spirit.

So I don't see how fixing some functionality in BM or advertising more boinc on social networks, where everyone focuses on arguments as mentioned above, could produce a "general interest" and a renewal of the boinc spirit. But I may be wrong, and I go as little as I can on SNs.

I recommend the interesting reading of the 2023 boinc census results - with just over half the participants of the previous census (and all of them already knew about boinc).

So maybe I'm all wrong but it's a simple fact that knowledge and interest in boinc has been constantly falling down for a serious number of years now, this hasn't started last but probably 15 years ago (David Anderson clearly commented this).

So personally I'll continue to crunch for boinc as long as I can, with or without FB (and I find strange the remarks above about German people "quitting boinc because of FB stopping" ?) but I don't have big global hopes around it. And l'AF forum is still alive, clearly not vivid as it was in the past, but alive


(*) I'm not talking about the excessive well-being of the small minority who want you to believe that you too can be in it, precisely so that they can continue to be in it, and not you, but this is a different topic

JeromeC wrote: ↑Sat Dec 30, 2023 1:50 pm
SPT : this is a "non project", some kind of Russian joke... working, not working, ended, not ended... not even recorded by all boinc statistical sites AFAIK.

This is not a joke.
This is an unsuccessful attempt to restore the stopped BOINC project T. Brada Experimental Grid.
The launch of the BOINC SPT project was carried out by an inexperienced person.
Gross mistakes were made.
As a result, the project does not work correctly.
I recommended stopping the project, but the administrator chose to continue.

I am reaching out to anyone who has experience with BOINC.
I need help starting a new project.
I have several algorithms that are different from the algorithm running in the BOINC SPT project.
One of these algorithms was tested in the BOINC project Gerasim@Home.

I have a small server (20 cores), this server was given to me by a colleague for remote control.
If you can try running a BOINC project on this server, please write to me.
I don't know if such a small server is suitable for a BOINC project.
But you can try the project for a small number of crunchers (by invitation).

By the way, there is a tuple competition going on right now
https://primesmagicgames.altervista.org ... k-tuple-2/

I invite everyone to the competition!
On the specified page you will find my e-mail.

Reminder : here we are in the Formula Boinc forum.