Sprints 2025 - timetable and projects selected

[zombie67]

This thread is sounding just like a race director's telephone line.

If you hit the throttle too hard you get wheel spin and oversteer which puts you back a few places, is that the driver's fault or the race track?

FB is mimicking F1 well. I bet fail2ban has never been called a "wet track" before.

Wait...are you putting the blame on the participants? You are saying that because some people tried to get work, and were banned, it is their fault?

Your analogy is not good. This is like a horse rase where half of the gates did not open until the other half of the horses are already 1/3 around the track when the rest of the gates opened. Pretty sure that the race results would not stand in that case.

But this is not up to me, and just my opinion.

[Tamagoch]

Maybe we need to set a clear list of issues when a sprint should be cancelled. But, of course, these issues should be really global. As far as it goes 20 from 24 users from L1 running with some results. So, for example, if this number gets too low, there could be a clear reason for a cancellation. But not this time, I suppose.

This competition gets too complicated and may need a book of regulations with a size comparable to that in F1. In this scenario we need a governing body like FIA to work this out. Timbo alone should not do it. Yes, he is a race director, not a series boss and not a federation president.

It's not only Asteroids@home issue. Every sprint, almost every one, I'm spending nearly an hour too feed all my hosts and not always it goes smoothly. I do it manually, one by one, with a starting bunker of 0.1 and I hope it helps to keep a competition fair, from my side at least. But it always goes wrong after first 5-10 minutes of a sprint when guys with huge farms hitting 'Allow new tasks' button with a bunker of 2-3 days. I don't blame exactly you, zombie67, we may all be victims here. It is just how it works. We're making a good stress-testing for the project maintainers and it's only up to them to use it properly.

So, everytime I'm sitting without tasks I think: 'At least I'm saving money on elecricity...'
Let's remain positive, keep crunching and having fun.

[JeromeC]

I clearly understand the general frustration and why some of you were not lucky against other, but you must also all seriously put the "boinc activity" into perspective, I fully agree with Tamagoch latest message : boinc IS A HOBBY, mainly maintained by unpaid volunteers, this is 100% true for all boinc statistical website (like FB) and community forums, and very true for many of the boinc projects themselves, running after tiny bits of money to make the things living over personal family time.

Some of the projects are very dynamically maintained and vivid, and some other much less. The discord discussion posted by Tim above is very interesting : the guy in charge was not even looking at the messages, after the call he reacted quickly, and he even switched off fail2ban which in my views is even against security principles (personally I would never deactivate fail2ban on my small linux host, there are zillions of IP trying to get in all day long, and for reasonable access attempts the ban is always temporary), but he did just to please the boinc / sprint community with a short term measure.

There is a very reduced list of active project, some of them can't handle "a sprint pressure" at all, and we are overall a relatively limited number of people participating to the feast : this is a niche activity.

So let's stay patient and reasonable and let's not start killing each other because the sprint has a hard start...

And about "cancelling the sprint" : personally I think it would be frankly exaggerated considering what's really at stake, and since Tim is putting all his personal time and energy to make live this FB thingy, I think this should be entirely his decision.

[UBT - Timbo]

Link?

Link removed, as requested by Asteroids admin.

regards
Tim

[UBT - Timbo]

Hi all

Thanks to everyone for your patience with regards to the latest Sprint.

I would like to cover a few things, which members may be aware of and others may not be.

1) I make a routine check of available projects to see which ones should NOT be included in the selection process, if they have few tasks available or they are having any server issues.

2) I then message the relevant project admin(s) before a Sprint - usually 3 or 4 days in advance but sometimes 2 days. This is so that they can advise of any issues and I can then initiate a change and appoint an alternative project, say a few hours before the official start.

3) Some admins do not reply So, I then have to check if the tasks are available in quantity and make a decision based on previous Sprints whether to go ahead. Sometimes, this is not easy to do, as potentially many projects might not have lots of tasks available so, the 1st choice project might still be the best choice, even without project admin(s) acknowledging the Sprint.

4) And ultimately, even with the best laid plans, things *can* go wrong. In this case, about 24 hours after the Sprint started, an admin was able to make some changes and this should have helped those members and teams taking part, who found they had connection issues - as there are still 3 more days to go, so there is still time to boost their position in the relevant League tables.

5) In general the overall success of the Sprints, over the years it has been running, is both a bonus and a curse to projects; they can get lots of benefits by having many completed work units returned in a short period of time, but the downside is that their servers can be overwhelmed, either due to bandwidth restrictions or simply overloading the server CPUs by having to deal with hundred, even thousands of connections. These issues have been known about for a long time and this is why some projects are excluded from Sprints and also why the announcement of the chosen project is made at the start time. This then shortens the window during which times the servers might be overly stressed. And hopefully projects are better able to cope with the demands placed upon them.

6) As far as cancelling a Sprint, I think this should only take place if there are NO stats being collected at all, (due to a faulty stats collection function on the FB website), or there is an unexpected and extended downtime from the selected project, both of which affects everyone. In this case, many could download tasks, as could be seen by the stats displayed on the FB website and a few members (and their large farm of hosts) could not, due to the project trying to protect itself from any possible DDOS attack.

7) As has been said, this is a hobby and we are all doing our bit to help the projects by volunteering our resources. And it means that all the chosen Marathon projects can be supported and 24 times a year specific projects are selected, and projects can get a helpful boost during these times.

regards
Tim

[JeromeC]

Amen

[UBT - wbiz]

Asteroids is out of tasks now with an extraordinary amount of tasks in progress. I hope this isn't somebody maliciously downloading all the tasks as I have seen happen before during sprints - computer full of tasks but tasks completed and no credit ever recorded on that computer.

[Tamagoch]

Asteroids is out of tasks now with an extraordinary amount of tasks in progress. I hope this isn't somebody maliciously downloading all the tasks as I have seen happen before during sprints - computer full of tasks but tasks completed and no credit ever recorded on that computer.

We shall see how much tasks would be cancelled and returned to a status page after the sprint.

I knew it's coming and have reduced bunker size on all machines this morning. It would be a shame to know I'm doing this 'fair play' in solitude.

[UBT - wbiz]

Asteroids is out of tasks now with an extraordinary amount of tasks in progress. I hope this isn't somebody maliciously downloading all the tasks as I have seen happen before during sprints - computer full of tasks but tasks completed and no credit ever recorded on that computer.

We shall see how much tasks would be cancelled and returned to a status page after the sprint.

I knew it's coming and have reduced bunker size on all machines this morning. It would be a shame to know I'm doing this 'fair play' in solitude.

I downloaded enough to the end of the sprint in the last 24 hours, it was amazingly accurate, not all projects are that accurate. The nefarious downloaders normally let the tasks time out rather than abort them, I presume they just wipe the computer, possibly a repeated number of times for those projects that limit number of downloads.

My unvalidated tasks increased about 60% in the last 24 hours of the sprint, it had been pretty stable up to then.

[UBT - Timbo]

I downloaded enough to the end of the sprint in the last 24 hours, it was amazingly accurate, not all projects are that accurate. The nefarious downloaders normally let the tasks time out rather than abort them, I presume they just wipe the computer, possibly a repeated number of times for those projects that limit number of downloads.

My unvalidated tasks increased about 60% in the last 24 hours of the sprint, it had been pretty stable up to then.

Hi

It is even possible that some may "tweak" the BOINC Manager configuration settings so as to increase the number of CPUs reported by BOINC Manager to the project and hence these hosts could download more tasks instead of being limited by the BOINC server settings

Hopefully all project admins can do something about this and reduce stress on the project servers by these saboteurs who wish to cause disruption to genuine volunteers on any project whether it is a Challenge, a Sprint or some other event?

regards
Tim