Formula Boinc Sprints 2018

[StefanR5R]

I had a suggestion. It takes forever to get new accounts approved, so if you want to suggest something else I can post it.

Regarding prior notification of project admins, here is an example of a simple routine:

• A few weeks before a sprint, select three projects randomly. This can be done on a relaxed schedule, but not too late. (Maybe a bit more than a "few" weeks are advisable.)
  
• Send an e-mail to the administrators of each of the three projects, asking whether it would be convenient to hold a contest at the given time frame. (Also tell the admins how much contributor capacity was seen at recent sprints, and that their project is one out of three taken into consideration this time.)
  
• As soon as the first positive reply is received, select this project, inform the admin of the choice and thank in advance, and send a friendly cancellation note to the other two admins. (These other two choices could be elected again for a subsequent sprint.)
  
• If only negative replies are received, cancel the sprint.
  
• If no replies are received at all, either cancel the sprint, or select one of the projects anyway, and state in the sprint announcement that no contact to the project admin could be established prior to the sprint.

(copied from the Asteroids forum, slightly edited)

--------
If the idea gets accepted that sprints could be canceled when there is no response from admins, or at least in case of negative responses, I suspect quite a few sprints in a year could end up canceled. Which would be disappointing to prospective sprinters. But sprint projects without work, too slow work generators, a defunct validator, an acoustic coupler for internet connection, ... are a bit of a disappointment too.

 

[ao_ika_red]

Not sure if you guys have seen this, but Seb is asking for suggestions on what to do about Sprints in 2019.

http://formula-boinc.org/forum/viewtopic.php?f=15&t=128

Wow, what's wrong with this marsinph user?

 

[lane42]

@tt
That's the way I get with no sunday morning stats

 

[zzuupp]

I'm lazy. I'll post my suggestion here.

For a BOINC project admin, ask for their approval at the beginning of the season. Let them OPT IN.
I'm sure some smaller projects will attempt to hang on. If the admin(s) don't know, don't care and don't reply, then that project if off the list.

 

[zzuupp]

Wow, what's wrong with this marsinph user?

Apparently, someone forgot their meds. Or maybe doubled down.

 

[StefanR5R]

For a BOINC project admin, ask for their approval at the beginning of the season. Let them OPT IN.

A project like Asteroids@home would need to be notified in advance of their actual sprint. They can have enough tasks, but only if they prepare for it. There are other projects which work similarly.

Or take Universe@home as example: Late in 2017 - early 2018, they were down entirely. Then they set up a new beta application and let it run by users who agreed to run test applications in their web preferences. They had to work out some bugs. But not long later, Universe was fully up and participated in the Pentathlon, and did well there.

IOW project status isn't constant over an entire year. It may not make sense for project admins to agree to a sprint if they still don't get to know when the sprint will happen.

 

[Howdy]

OK, how about this:
Generate the project list for the whole year/ schedule in advance or maybe a month or 2 in advance.(still only releasing the chosen project as it is now) Contact said projects admin, inform them on date that the sprint would run. Allow the admin to opt in/out, set that project for that date if in/ choose another if they opt out or no reply. The only thing that would be counted on is the admins honesty in keeping it secret.

 

[ao_ika_red]

Apparently, someone forgot their meds. Or maybe doubled down.

That's really strong med he got then.

The only thing that would be counted on is the admins honesty in keeping it secret.

That would spark another controversy.

 

[Howdy]

That would spark another controversy.

Most certainly.................seems to be the "norm" these days.

 

[zzuupp]

Most certainly.................seems to be the "norm" these days.

Alas, that is true. Any excuse for drama makes it mandatory.

 

[TennesseeTony]

0500 UTC start time for the next two Sprints, then 0200 UTC for the next to last one. Then back to 'normal' at 2000 UTC for the very last one.

 

[crashtech]

0500 UTC will be 2300 MST for me on a day when I am waking up at 0530. Gonna be a long day...

 

[StefanR5R]

The purpose of the 24 h announcement period is that everybody can start on their own time. There is no written rule that each single one of the 1440 minutes lead time must be used for bunkering.

(I had plenty of sprints starting at 00:00 local time. Sometimes I stood up late and cut into my sleep. Sometimes I slept through the first hours of the announcement and managed to set up everything before leaving for work. Other times I missed most of day 0 and started just a few hours before the actual sprint.)

 

[crashtech]

The purpose of the 24 h announcement period is that everybody can start on their own time. There is no written rule that each single one of the 1440 minutes lead time must be used for bunkering.

(I had plenty of sprints starting at 00:00 local time. Sometimes I stood up late and cut into my sleep. Sometimes I slept through the first hours of the announcement and managed to set up everything before leaving for work. Other times I missed most of day 0 and started just a few hours before the actual sprint.)

Well, sure. I suppose I like to make up for my lack of finesse in configuring my systems by using every crude method available to me; maximizing bunker time is easy to understand and implement, if time consuming. We'll see if I can keep my eyelids propped open long enough...

 

[Orange Kid]

Wake up it's Universe@Home

 

[ao_ika_red]

Wake up it's Universe@Home

Optimised for linux

 

[StefanR5R]

Results of the Universe@home sprint in November 2017:

SETI.USA .............................. 23 M
Gridcoin .............................. 19 M
TeAm AnandTech ................. 9 M
Czech National Team ........... 8 M
​

Gridcoin will probably produce >30 M in the current sprint.

(edit) These league 1 teams registered for the 3rd week of the Thor challenge and may (or may not) be diverting CPUs to bunkering WCG now: Planet 3DNow!, SETI.Germany, SETI.USA, L'Alliance Francophone, Team China, BOINC.Italy.

 

[StefanR5R]

There is more variance in the run time of the current tasks than I remember from Pentathlon. I was just trying to figure out how many more tasks to download for the rest of the sprint, but it doesn't work out.

 

[TennesseeTony]

Optimised for linux

Well, it used to be, for sure. I am not so certain anymore, however.

 

[Orange Kid]

Universe@home: Day one is about done and we are in second.
The usual players seem to be absent from this sprint. Maybe they have yet to release their bunkers. Time will tell.

1 Gridcoin 25 ___9,625,333
2 TeAm AnandTech 18 ___4,536,000
3 BOINC@AUSTRALIA 15 ___2,034,000
4 Planet 3DNow! 12 ___1,970,666
5 BOINC@Poland 10 ___1,724,000
6 Czech National Team 8 ___1,096,666
7 Rechenkraft.net 6 ___1,037,333
8 Russia Team 4 ___994,666
9 Dutch Power Cows 2 ___562,000
10 Overclock.net 1 ___480,666

FreeDC is down.

 

[ao_ika_red]

Well, it used to be, for sure. I am not so certain anymore, however.

Well, thanks to unfortunate event which was occured when I was bunkering, I can tell that on linux it's much faster than on Windows.
My laptop's task completion time on linux was about 2 hours, and after I switched to Windows it needs about 6 hours to complete.

 

[biodoc]

There are WU that take much longer to complete. My single socket ivy Bridge now has a bunch of WU that take about 10 hours to complete on linux. Before, the same machine was getting tasks that took under 2 hours to finish.

 

[TennesseeTony]

I have compared the completed tasks between the two machines I listed.

There seems to be about 6 different groups of tasks:

• 1.4 hours
• 2.5 hours
  
• 3 hours
• 3.6 hours
• 5+ hours
• 10+ hours

Those times are for Win10, equivalent Linux times were maybe 10 minutes faster. I only saw 1.4 hours and 10+ hours on Linux. Otherwise the mix looked pretty even across those two machines. There were only 2 of the 10 hours tasks, so not enough to blame for the performance loss.

In a nutshell, I'd say run Linux if you can, but it is good to see that Windows is no longer so far behind (and in fact with these two machines, Windows is widening the lead even further).

 

[Howdy]

It would be nice if the points scaled higher with the completion times being longer instead of the set "devil number.67" but alas..........."it is what it is"

 

[StefanR5R]

First half of the sprint, top 10 teams of league 1:

Meanwhile, there are still 145,000 tasks in progress, which would be 97 million points after validation. Only a part of these will be reported and validated during the sprint, yet it could still bring about considerable changes in any of the leagues...

(server status graph courtesy @Kiska)

--------
(edit) Roughly 1 per 1000 tasks that I completed so far got stuck without making further progress. At least half of these were successfully completed after going through one or two suspend/ resume cycles.