Re: Formula BOINC 2025

[rimmer66]

looks like we will win the second sprint ,

[wbiz]

looks like we will win the second sprint ,

Easily, nobody can catch us now.

[UBT - Timbo]

Hi both

Looks like UBT did pretty well:

https://www.formula-boinc.org/sprints.p ... printid=28

Mind you: League 3 is made up of the weakest FB teams (from 2024) that have expressed a willingness to take part in the 2025 season.

And UBT might not do so well, once the mid-season promotions/relegations take place as the new Team for 2025, Atlantis, is doing pretty well in League 4. But then again they might be promoted to League 2, along with UBT, *IF* they continue on an upwards trajectory...AND assuming some teams in either of the higher leagues fail to support FB and they get relegated.

regards
Tim

[wbiz]

With the existing members of League 2, we would have come 7th in the first sprint and 6th in the second which a fair position, we would be 7th in the overall L2 sprints but yes it probably will get tougher if the leagues are updated.

Difficult to tell where we would be at the moment in L2, there is a different spread of projects in different leagues, when we changed from one league to multiple leagues, my position in L2 was initially the same (6th) as the overall position was, I still trying to get my head round that, it is very counterintuitive.

I'm hoping einstein doesn't pop up in the sprints, its usually one of our weakest sprint projects.

We have a very wide spread of marathon projects which reduces risk of getting wiped out, this is pretty much my strategy at the moment, I've learnt from past mistakes it doesn't pay to put your eggs in only a few baskets.

We are 10th overall on boincgames out of 28 active teams, which is very good. None of our team did any tasks on their first sprint (milkyway), partly a mistake on my part, I meant to do one or two tasks just to get the bare minimum 200-ish points for the team and myself. This will probably cost us a couple of places shortly.

[wbiz]

Current sprint is NFS, ends on Monday 7th April 0700 UTC (0800 BST)

[wbiz]

Current sprint is NFS, ends on Monday 7th April 0700 UTC (0800 BST)

Team came first again!

[wbiz]

Current sprint is Asteroids, finishes Sunday 13th April at 1700 UTC (1800 BST)

[rimmer66]

looks like we will run away with this one as well .

[wbiz]

looks like we will run away with this one as well .

Aye, we have now come first in all four sprints so far.

I struggled with this one, I eventually threw the kitchen sink at it, 62 cores 98 threads, I was doing ok for 24 hours or more, peaked at 2nd but then got overtaken many times.

[wbiz]

Current sprint is Einstein which ends Sunday 20th April 19:00 UTC (20:00 BST).

Not one for me unfortunately.

[wbiz]

We came third in the last sprint but we are still a very comfortable first overall. Next sprint is the 1st May at 22:00 UTC (23:00 BST)

[rimmer66]

lets hope its something you can crunch , we were a very close third.

[wbiz]

lets hope its something you can crunch , we were a very close third.

Whoops! Normally my contribution to Einstein sprints make no difference, this time it may have done but I suspect if I'd joined in then TFF would have upped their games so the UBT result would be unaffected.

TBH I can't remember the exact reason for my aversion to Einstein, possibly the accumulation of it being predominantly GPU, ridiculously long tasks, two user validation, validations taking forever (just had one take 16 days!) and a non-standard website.

3 days out would have cost me some places on marathon projects so I only feel a little bit guilty

[wbiz]

Current sprint is SRBase which ends on Sunday 4th May 2200 UTC (2300 BST).

[rimmer66]

i would like to post

[rimmer66]

or post an edit
sorry i can't crunch to my potential

[rimmer66]

i live in a studio flat and the gpu fan blasting constantly is to much ,
sorry

[rimmer66]

sorry about the format ,
i was an arsehole to timbo a while ago , i think i may be still a bi blocked.

[wbiz]

i live in a studio flat and the gpu fan blasting constantly is to much ,
sorry

We came third in the sprint and we are still a very safe 1st position overall.

We haven't been a strong team on GPU sprints for a long time so I normally accept our fate but this time there was little activity from other teams in our league. I mostly used three embedded GPU's, the rest were CPUs. I wasn't going to be able to stay in the user points and the team had a pretty safe third position so I stopped sprinting very early, SRBase isn't one of my target projects this year.

[UBT - Timbo]

sorry about the format ,
i was an arsehole to timbo a while ago , i think i may be still a bi blocked.

Hi

You are not blocked at all....not sure how you thought that, as clearly you can post here, without my intervention,

And I tend to let "sleeping dogs lie" so if you were an arsehole, it doesn't bother me ! In fact I cannot even remember such an event...so I doubt it was a serious matter

regards
Tim

[UBT - Timbo]

We came third in the sprint and we are still a very safe 1st position overall.

Hiya

I think we came 4th, beaten by Switzerland, by just 400 BOINC credits:

https://www.formula-boinc.org/sprints.p ... printid=32

Sadly, I could only start crunching on Sunday, as I could not access the PC in my office as the remote control software was not connecting (as my office PC had "locked up") and I had to go there to do an "end of year" stock take, so I was able to reboot the PC and get it running the Sprint project.

The shorter running tasks (which i think were GPU - short v0.47) my host crunched only earned 300 credits each, while the longer running tasks (which I think were CPU - average2 v0.47) earned 500 credits each. But sadly, 10 of the tasks I had ended AFTER the Sprint finished..so they didn't count

We haven't been a strong team on GPU sprints for a long time so I normally accept our fate but this time there was little activity from other teams in our league. I mostly used three embedded GPU's, the rest were CPUs. I wasn't going to be able to stay in the user points and the team had a pretty safe third position so I stopped sprinting very early, SRBase isn't one of my target projects this year.

The next Sprint is not due to start for about 10 days, so, maybe a CPU-only project might be selected...or maybe not... ...

regards
Tim

[wbiz]

We came third in the sprint and we are still a very safe 1st position overall.

Hiya

I think we came 4th, beaten by Switzerland, by just 400 BOINC credits:

Weird, I was watching their progress and could have sworn they hadn't make it to 3rd a while after 11pm, and its now 400 credits difference.

Did the anti-bunkering trial fail? The last two sprints there was some heavy bunkering.

We still have the invisible man competing in user-L2, I don't mind the username missing but when you can't see the credits of the person you are chasing you have no idea if you are gaining or just throwing computer time away.

[UBT - Timbo]

Hiya

I think we came 4th, beaten by Switzerland, by just 400 BOINC credits:

Weird, I was watching their progress and could have sworn they hadn't make it to 3rd a while after 11pm, and its now 400 credits difference.

Hi - I think I mis-read the final stats as the 0 and the 6 look very similar...so, yes we got beat by 400 credits. (I've corrected my post above)

Did the anti-bunkering trial fail? The last two sprints there was some heavy bunkering.

The only "anti-bunkering" function is that the project is no long announced in advance. There is nothing else that can be done about Members/Teams "speculating" as to which project might be chosen - and I doubt that the project admin would make any announcement BEFORE the start time.

However, efforts are made (in advance) to ensure that the selected Sprint project will have enough tasks available, but some Members/Teams could easily have enough hosts to be crunching a few projects and once the announcement is made, they can easily stop accepting tasks from other projects.

We still have the invisible man competing in user-L2, I don't mind the username missing but when you can't see the credits of the person you are chasing you have no idea if you are gaining or just throwing computer time away.

I've only just seen your post from early April (on the FB forum) about the missing member(s). So I'll advise Sebastien and see what he has to say about it (or to make any corrections).

regards
Tim

[wbiz]

The only "anti-bunkering" function is that the project is no long announced in advance. There is nothing else that can be done about Members/Teams "speculating" as to which project might be chosen - and I doubt that the project admin would make any announcement BEFORE the start time.

I may have dreamt it or totally misunderstood it and it might not have been for the purposes of anti-bunkering but I thought Sebastian did some work on only incorporating credits on tasks that were issued after the start of the sprint. That data is available but it may create a significantly heavier workload on servers.


The invisible man is currently in the L2 marathon positions:-

10th Amicable
15th Milkyway
9th Moo! Wrapper
16th NFS
8th ODLK
7th ODLK1
18th Primegrid
9th RakeSearch
14th Rosetta
12th Sidock
15th SRBase
10th Ramanujam
10th YoYo

[wbiz]

Current sprint is Ramanujan which ends on Sunday 18th May 1500 UTC (1600 BST)

EDIT: Got my GMT and BST mixed up doh!

[UBT - Mikee]

Current sprint is Ramanujan which ends on Sunday 18th May 1500 UTC (1600 GMT)

Hi

Once again late to the party but as the site is currently down I didn't know what the sprint was - luckily you mentioned it so I'm here but only running intermittently.

Mike

[wbiz]

Current sprint is Ramanujan which ends on Sunday 18th May 1500 UTC (1600 GMT)

Hi

Once again late to the party but as the site is currently down I didn't know what the sprint was - luckily you mentioned it so I'm here but only running intermittently.

Mike

Welcome mikee, I see you have arrived straight into validation heaven, hopefully more of your tasks will tunnel their way out.

[UBT - Mikee]

'Validation Heaven'?

Not the most precise description but I know what you mean.

Been a whlle since I had to go through wingmen to get credits - thought we had got past that but there we go...

Mike

later...

Where's "Marvin The Dalek" avatar gone??? Or did he go in the Forum cull?

[wbiz]

Current sprint is Gerasim which ends on Sunday 25th May 1500 UTC (1600 BST)

[wbiz]

A comfortable first in the sprint for the team

[wbiz]

Current sprint is Numberfields which ends on Sunday 1st June 1500 UTC (1600 BST)

[UBT - Timbo]

Where's "Marvin The Dalek" avatar gone??? Or did he go in the Forum cull?

HI Mikee

There's been no "forum cull" as such so, I have no idea why the avatar is missing...let me investigate further.

regards
Tim

[UBT - Timbo]

Where's "Marvin The Dalek" avatar gone??? Or did he go in the Forum cull?

Hi again

I have NO CLUE as to what's happened on the graphics side of things - even the main logo (in the big blue box above) has vanished as well !!

I have adjusted one or two things and my own avatar has re-surfaced...but others, such as yours, have not returned.

I'll try a couple of other things, but it's got me scratching my head a bit as to what's happened as I've not changed anything in the Admin Control Panel.

regards
Tim

[wbiz]

A very respectable 3rd place in the Numberfields sprint, tough sprint!

[wbiz]

Current sprint is Asteroids, finishes on Sunday 29th June at 1500 UTC (1600 GMT)

[wbiz]

Another sprint first place for the team. One more sprint before promotion to the 2nd League

[wbiz]

Current sprint is Gerasim which ends Sunday 6th July 1600 UTC (1700 BST)

[wbiz]

And the team came first again in the sprint, that's 7 firsts out of 12 sprints.

[wbiz]

The team has been promoted to League 2 .... and are now top of L2 and a full hat trick again:- sprints, marathon and overall. I didn't see that coming!

[wbiz]

Current sprint is SRBase which ends Sunday 27th July 1500 UTC (1600 BST)

Note if you have updated to latest Boinc-Client on Windows and configured WSL2 and Podman (or Docker I presume), this project is making use of it so will run under linux on a windows machine.

Not so sure what is happening on linux with podman as yet, this is getting far too complicated for me.

[UBT - Timbo]

The team has been promoted to League 2 .... and are now top of L2 and a full hat trick again:- sprints, marathon and overall. I didn't see that coming!

Hiya

It's not much of a surprise when you consider you have earned most of them:

Code: Select all

		Placing	Name		FB Points
Teams League 2:
Totals:		1	UK BOINC Team		645
Marathon:	1	UK BOINC Team		387
Sprints: 	1	UK BOINC Team		258

Members League 2:
Totals: 	4	UBT - wbiz		340
Marathon:	4	UBT - wbiz		248
Sprints: 	9	UBT - wbiz		92

Members League 3:
Totals: 	21	UBT - Timbo		33
Marathon:	21	UBT - Timbo		5
Sprints: 	19	UBT - Timbo		28

Members League 4:
Totals: 	19	Stephen Hunter		80
Marathon:	18	Stephen Hunter		68
Sprints: 	19=	Stephen Hunter		12

regards
Tim

[wbiz]

The team has been promoted to League 2 .... and are now top of L2 and a full hat trick again:- sprints, marathon and overall. I didn't see that coming!

Hiya

It's not much of a surprise when you consider you have earned most of them:

That is extremely misleading, the FB Team points and FB Member points aren't related.

Most team marathon project points are very little to do with me, the wider membership are doing an excellent job.

I'm very concerned about where boinc is heading using containers, with Podman the sharing of some resources is static (eg swap space, you configure how much swap every container gets, I presume that is a share of the global swap?), I don't know if that applies to Docker as well, being daemon based Docker might have dynamic allocation?

I think boinc will lose a lot of users as configuration gets more complicated and increased limitations such as static resource allocations. We have had enough trouble with some projects running out of memory without doubling up on that problem.

[UBT - Timbo]

That is extremely misleading, the FB Team points and FB Member points aren't related.

Most team marathon project points are very little to do with me, the wider membership are doing an excellent job.

Hiya

The FB Team points are based on every team members BOINC credits, whether for the Marathon or for any Sprint.

And FB Member points are ONLY counted for those who have registered as a individual member

The actual breakdown of FB points (as of my stats provided before) is as follows:

Totals: UK BOINC Team 645 of which: 340 (wbiz) + 33 (timbo) + 80 (SH) = 453 and rest of team = 192
Marathon: UK BOINC Team 387 of which: 248 (wbiz) + 5 (timbo) + 68 (sh) = 321 and rest of team = 66
Sprints: UK BOINC Team 258 of which: 92 (wbiz) + 28 (timbo) + 12 (sh) = 132 and rest of team = 126

I'm very concerned about where boinc is heading using containers, with Podman the sharing of some resources is static (eg swap space, you configure how much swap every container gets, I presume that is a share of the global swap?), I don't know if that applies to Docker as well, being daemon based Docker might have dynamic allocation?

I think boinc will lose a lot of users as configuration gets more complicated and increased limitations such as static resource allocations. We have had enough trouble with some projects running out of memory without doubling up on that problem.

Since David Anderson stopped being actively involved in BOINC Manager development (and I presume BOINC Server development) so it's been left to others and in all seriousness I do not think they have a clue what they are doing.

(See what David thinks of BOINC, as of 26th Jan 2022, and (c) 2025 here: https://continuum-hypothesis.com/boinc_history.php

David's summary about BOINC says this:

The decline of BOINC
Pretty early on - maybe 2007 - it became clear to me that BOINC had plateaued. It has gradually declined ever since:

After an initial flurry of projects, there has been maybe one new project a year, and they tend to be fringe things like exhaustive-search Math stuff, rather than mainstream science.

I didn't keep records (I should have), but the number of volunteered computers also started declining. Except for a few bumps (like the onset of COVID) it has continued to do so. From a high of maybe 700K, it's down to 60K, and falling.

As of early 2022, the "BOINC Community" is down to a few diehards. The level of open-source software contribution is small.

The term "volunteer computing" hasn't appeared in a CS conference announcement or an NSF Call for Proposals for 15 years.
BOINC has received no coverage in the mass media for years, and it has negligible presence on social media.

From 2007 onward, I tried to reverse the decline - first by making the software more powerful and general, then by building bridges with the HPC mainstream. Eventually I realized that the free-market model was a fundamental problem.

In an ideal world, BOINC Manager should have evolved to include more user functionality (like creating/editing project XML config files), more ability to assign specific processors to specific projects in any single host (so for a 16 core CPU, allow (say) 2 for CPDN, 4 for PrimeGrid and 8 for Rosetta and leave 1 for GPU control/functions).

Likewise the ability to remote control any named hosts within BOINC Manager would be good, like "BOINC Tasks" or "BOINC View" and to control how much resources are used with "TThrottle". (See here for more links: https://setiathome.berkeley.edu/download_network.php

But no-one has done much about enhancing BOINC Manager to take all these other options into account.

regards
Tim

[wbiz]

That is extremely misleading, the FB Team points and FB Member points aren't related.

Most team marathon project points are very little to do with me, the wider membership are doing an excellent job.

The actual breakdown of FB points (as of my stats provided before) is as follows:

Totals: UK BOINC Team 645 of which: 340 (wbiz) + 33 (timbo) + 80 (SH) = 453 and rest of team = 192
Marathon: UK BOINC Team 387 of which: 248 (wbiz) + 5 (timbo) + 68 (sh) = 321 and rest of team = 66
Sprints: UK BOINC Team 258 of which: 92 (wbiz) + 28 (timbo) + 12 (sh) = 132 and rest of team = 126

The team marathon project points are calculated from the summation of user project credits not on summation of FB points, which in turn are summated for the overall team marathon position.

Look at Primegrid, Einstein, WCG, Amicable etc etc where the team has a good standing but I am only watering the wind compared to the project credits that are used to calculate the FB points. I only have a significant contribution to a small number of more niche or intermittent projects because that is the tactical competition I play due to lack of raw power. The overall team marathon success is certainly not down to me and probably not even me, you and SH added together (if you want, I will knock some figures together).

In an ideal world, BOINC Manager should have evolved to include more user functionality (like creating/editing project XML config files), more ability to assign specific processors to specific projects in any single host (so for a 16 core CPU, allow (say) 2 for CPDN, 4 for PrimeGrid and 8 for Rosetta and leave 1 for GPU control/functions).

Likewise the ability to remote control any named hosts within BOINC Manager would be good, like "BOINC Tasks" or "BOINC View" and to control how much resources are used with "TThrottle". (See here for more links: https://setiathome.berkeley.edu/download_network.php

Yes, Boinc manager was mentioned at stonehenge, it desperately needs improved ability. I use it remotely on the windows machines over VNC but that will end shortly now that Windows has a command line editor.

I use a single bointui to manage all my projects on Windows and Linux. I use a little shell script to remotely edit the xml based settings over SSH, I never got round to including Windows in that because of a lack of command line editor, it would have meant downloading the file editing locally and uploading again which I decided against. Now "edit" is available (again) on windows I can remote edit over SSH like I do with the linux machines.

[wbiz]

That is extremely misleading, the FB Team points and FB Member points aren't related.

Most team marathon project points are very little to do with me, the wider membership are doing an excellent job.

Hiya

The FB Team points are based on every team members BOINC credits, whether for the Marathon or for any Sprint.

And FB Member points are ONLY counted for those who have registered as a individual member

The actual breakdown of FB points (as of my stats provided before) is as follows:

Totals: UK BOINC Team 645 of which: 340 (wbiz) + 33 (timbo) + 80 (SH) = 453 and rest of team = 192
Marathon: UK BOINC Team 387 of which: 248 (wbiz) + 5 (timbo) + 68 (sh) = 321 and rest of team = 66
Sprints: UK BOINC Team 258 of which: 92 (wbiz) + 28 (timbo) + 12 (sh) = 132 and rest of team = 126

Here are the marathon figures as date of posting, in 16 of the 27 projects, other members contributed more than the sum of the gang-of-three you mentioned. So as I correctly predicted, other members contributed more to our overall marathon score than the gang-of-three.

The anomaly with ODLK25 is because I registered with ODLK25 a short time before the team did, I couldn't be bothered bodging a correction. I have removed Universe@Home because that is zero.

[UBT - Timbo]

Here are the marathon figures as date of posting, in 16 of the 27 projects, other members contributed more than the sum of the gang-of-three you mentioned. So as I correctly predicted, other members contributed more to our overall marathon score than the gang-of-three.

The anomaly with ODLK25 is because I registered with ODLK25 a short time before the team did, I couldn't be bothered bodging a correction. I have removed Universe@Home because that is zero.

Hiya

Thanks for your further post.

I did try to reply previously, TWICE, but both times, when I hit "submit", I got turfed out of the forum and back to the login page - and of course I lost the messages I had composed.

I am not sure why this happened and hopefully this reply will be OK.

Yes, I think I made a faux pas and incorrectly included the "gang of three" individual FB points as part of the Team total FB points. I think my brain was a bit fogged up when I did that. So sorry for the error.

regards
Tim

[wbiz]

Current Sprint is NFS which ends Sunday 3rd August at 1500 UTC (1600 BST)