FB Sprint 2026 Notifications

[UBT - Timbo]

Hi all

The 1st Sprint starts next week !!

2026-03-05 06:00 UTC - 2026-03-08 06:00 UTC

and the first project has now been randomly selected.

Once their admin confirms there are no planed maintenance periods for the above dates, then it will be "locked in".

Good luck to everyone for 2026 !

regards
Tim

[UBT - Timbo]

Hi all

The 1st Sprint project, randomly selected, of 2026 is PrimeGrid

Please note that PrimeGrid is starting a "Telephone Day" Challenge on 7th March, that only involves the "Seventeen or Bust" (SoB) sub-project.

These can be very long work units, so do not crunch these for the Sprint - choose other ones instead for the Sprint - as the SoB tasks might not finish until after the Sprint ends - and besides which, tasks downloaded BEFORE the start of the Challenge will not count towards your Challenge stats !

Good luck everyone !

regards
Tim

[UBT - Timbo]

Hi all

The 2nd Sprint starts tomorrow !!

2026-03-12 09:00 UTC - 2026-03-15 09:00 UTC

and a project has already been randomly selected.

https://www.formulaboinc.org/sprints.py ... printid=54

Their admin has also confirmed that tasks will be available, so all should be well, during the weekend.

The selected project is Asteroids@home

regards
Tim

[csbyseti]

Nice to see that the Asteroids@home Admins have saved all errors from last year.

Same none working Boinc server as every year.

csbyseti

[JeromeC]

It's very reassuring; we know where we stand.

jeu. 12 mars 20:55:59 2026 | | Project communication failed: attempting access to reference site
jeu. 12 mars 20:55:59 2026 | Asteroids@home | Scheduler list fetch from https://asteroidsathome.net/boinc/ failed: transient HTTP error
jeu. 12 mars 20:56:00 2026 | | Internet access OK - project servers may be temporarily down.

[UBT - Timbo]

Hi both

It is entirely possible that the errors you are experiencing could be caused by many factors - and sadly BOINC Manager does not help by providing inadequate responses whenever something doesn't quite work.

A loss of connectivity (to the Asteroid server) can easily occur when too many connections are being made from multiple hosts, anywhere in the world.

And if you recall in the past, Asteroids was hit by a DDOS attack that started when a previous Sprint got under way and the perpetrators hid themselves behind one (or more) VPNs.

Personally speaking, I've had very few issues** getting tasks and as long as you keep a small but acceptable "bunker" of tasks, then BOINC Manager will always keep re-trying to obtain more work units.

Hopefully, the glitch you've noticed is temporary and has not affected your crunching?

regards
Tim

1st UPDATE:

** The only times I've had connection issues, during this Sprint, were:
11/03/2026 13:17:57 | | Project communication failed: attempting access to reference site
11/03/2026 13:17:58 | | Internet access OK - project servers may be temporarily down.
12/03/2026 13:18:05 | | Project communication failed: attempting access to reference site
12/03/2026 13:18:07 | | Internet access OK - project servers may be temporarily down.
13/03/2026 13:18:19 | | Project communication failed: attempting access to reference site
13/03/2026 13:18:26 | | Internet access OK - project servers may be temporarily down.

and spot how regular this has happened - almost exactly 24 hours apart !!

and there have been times when no tasks were available:
12/03/2026 12:08:17 | Asteroids@home | work fetch resumed by user
12/03/2026 12:08:45 | Asteroids@home | Project has no tasks available
12/03/2026 12:08:56 | Asteroids@home | update requested by user
12/03/2026 12:09:05 | Asteroids@home | Project has no tasks available

12/03/2026 18:00:13 | Asteroids@home | Project has no tasks available
12/03/2026 21:05:23 | Asteroids@home | Project has no tasks available
12/03/2026 22:21:34 | Asteroids@home | Project has no tasks available

13/03/2026 08:01:06 | Asteroids@home | Project has no tasks available
13/03/2026 10:20:33 | Asteroids@home | Project has no tasks available
13/03/2026 10:22:40 | Asteroids@home | Project has no tasks available

2nd UPDATE:

13/03/2026 13:33:21 | Asteroids@home | Project has no tasks available
13/03/2026 13:55:40 | Asteroids@home | Project has no tasks available
13/03/2026 13:57:53 | Asteroids@home | Project has no tasks available
13/03/2026 14:00:05 | Asteroids@home | Project has no tasks available
13/03/2026 14:02:18 | Asteroids@home | Project has no tasks available
13/03/2026 14:07:00 | Asteroids@home | Project has no tasks available
13/03/2026 14:15:03 | Asteroids@home | Project has no tasks available
13/03/2026 14:17:35 | Asteroids@home | Project has no tasks available
13/03/2026 14:21:33 | Asteroids@home | Project has no tasks available
13/03/2026 17:31:38 | Asteroids@home | Project has no tasks available
13/03/2026 17:37:47 | Asteroids@home | Project has no tasks available
13/03/2026 17:57:16 | Asteroids@home | Project has no tasks available
13/03/2026 17:59:54 | Asteroids@home | Project has no tasks available
13/03/2026 18:02:09 | Asteroids@home | Project has no tasks available
13/03/2026 18:04:22 | Asteroids@home | Project has no tasks available
13/03/2026 18:06:35 | Asteroids@home | Project has no tasks available
13/03/2026 18:08:48 | Asteroids@home | Project has no tasks available
13/03/2026 18:16:04 | Asteroids@home | Project has no tasks available

But my bunker does not appear to have run out of tasks as uploads are still happening at regular intervals - if there were big gaps between successive uploads, one might expect that my bunker was empty but that does not appear to be happening.

[UBT - wbiz]

I had the tale of two cities.

One computer nicely filled up its cache and kept it topped up regularly. Another identical computer barely got any cache and ran out of work for a few hours. The rest of my computers bumbled along between those two extremes.

One particularly annoying aspect is the time estimation for my AMD Windows computers is completely wrong. It estimates 38 mins when tasks are actually 75 minutes, I can easily end up with loads of unfinished tasks left at the end of the sprint, which also messes up other peoples validation. My Intel and ARM linux computers have correct estimates.

[UBT - Timbo]

Hi all

BREAKING NEWS

It has been reported on the BBC website that the Bahrain and Saudi F1 Grand Prixs are very likely going to be cancelled and the use of other tracks has been discounted.

https://www.bbc.co.uk/sport/formula1/ar ... 75vw1lwjzo

Therefore, *if confirmed*, then there will only be 22 F1 races and not 24...and there will not be any F1 races in April as there'll be a 5 week gap between the Japanese GP from 27th-29th March and the Miami GP 1st-3rd May

So, any feelings for whether we should continue to hold Sprints on the two relevant weekends? Do we go ahead or cancel them?

Please feel free to let me know your opinions.

regards
Tim

[UBT - Timbo]

One particularly annoying aspect is the time estimation for my AMD Windows computers is completely wrong. It estimates 38 mins when tasks are actually 75 minutes, I can easily end up with loads of unfinished tasks left at the end of the sprint, which also messes up other peoples validation. My Intel and ARM linux computers have correct estimates.

Hiya

I've only got one PC running BOINC at present, and initially, the tasks were claimed to finish very quickly...but now it has settled down now and all my cache of tasks are reporting a duration time of about 2hr 40m (about 160 mins). And the validated tasks are around that duration.

So, I suspect the benchmarking on one of your hosts is a bit out of kilter?

But you can always adjust your cache to a smaller amount, as the admins have promised to make more tasks available, *if* FB Teams and Members are depleting the number of available tasks too quickly.

regards
Tim

[UBT - wbiz]

One particularly annoying aspect is the time estimation for my AMD Windows computers is completely wrong. It estimates 38 mins when tasks are actually 75 minutes, I can easily end up with loads of unfinished tasks left at the end of the sprint, which also messes up other peoples validation. My Intel and ARM linux computers have correct estimates.

Hiya

I've only got one PC running BOINC at present, and initially, the tasks were claimed to finish very quickly...but now it has settled down now and all my cache of tasks are reporting a duration time of about 2hr 40m (about 160 mins). And the validated tasks are around that duration.

So, I suspect the benchmarking on one of your hosts is a bit out of kilter?

But you can always adjust your cache to a smaller amount, as the admins have promised to make more tasks available, *if* FB Teams and Members are depleting the number of available tasks too quickly.

regards
Tim

Three identical computers doing the same thing, and they always have done on Asteroids and one other project (forgotten which). There doesn't appear to be any attempt for it to adjust to the correct time of around 75 mins all three are still estimating around 38 minutes. These are CPU tasks, they run day-in day-out on other projects with accurate estimates.

A quick sample of six tasks spaced apart shows the validators were all doing them in about 43 minutes apart from one who achieved 23 minutes, so I guess their averages are nearly 38 minutes but they are different CPUs and Asteroids shouldn't be imposing their average on my CPUs.

If I reduce my cache I risk running out of work as happened with one computer, supply of tasks is very erratic. The one that ran out didn't get any new tasks for about 6 hours. I do tail off my cache size as the sprint progresses to try and end at the end.