2020 Formula Boinc Sprints

[StefanR5R]

OK. Asteroids@home has 7464 tasks left ready to send. I switched my best computer back on, fetched ≈2000 tasks, set NNT, and let things run their course.

I meant to go on an electric diet during this short pause until Pentathlon. But my newest toy is so efficient that I should go this little extra step.

 

[TennesseeTony]

It is okay if you beat me to 50M......no, really. The Pentathlon will delay me, but my end goal is 100M.

 

[StefanR5R]

If I have ARM wingmen and you have x86 wingmen, you may still easily get there first. ;-)
.
.
.
Though I am beginning to suspect that you are wingman of most of my tasks.
BTW, how does host 635833 have 2335 tasks in progress? Communication errors, or modded client?

 

[Markfw]

If I have ARM wingmen and you have x86 wingmen, you may still easily get there first. ;-)
.
.
.
Though I am beginning to suspect that you are wingman of most of my tasks.
BTW, how does host 635833 have 2335 tasks in progress? Communication errors, or modded client?

Well, all tasks that are downloaded, and ready for processing or processing show as "in progress". At least for Rosetta and WCG for me.

 

[StefanR5R]

Tasks in progress are the sum of downloading, downloaded, running, suspended, uploading, and ready-to-report tasks.

When downloading, downloaded, and running tasks together are 1000 or more, the standard client does not request any more new tasks. Further, when one or more task of a project is suspended, the client does not request new tasks for this project. If a certain number of tasks are uploading to a project, the client does not request new tasks of this project either. If the client requests new work, it will also report any ready-to-report tasks, and so these ones are no longer in progress.

The upshot of all this is, a host with a standard client won't have much more than 1000 tasks of one and the same project in progress.

Exception: If there are failures in communication, particularly

• server considers work requests complete, client considers same requests failed,
• client considers reports of results complete, server considers same reports failed,

then the database at the server will hold more tasks in progress on that host than this host actually has in progress.

The above mentioned limit in the client which stops it from requesting work after it collected 1000 runnable tasks can be lifted by building a client from modified source code.

— The upshot: 2335 tasks in progress (on one host at one project) is really unusual and won't normally happen.

 

[TennesseeTony]

....
BTW, how does host 635833 have 2335 tasks in progress? Communication errors, or modded client?

Ah, I missed this the first time I read that. The answer is quite simple, really......uh, I have no idea. I doubt they would finish before timing out so I aborted most of them and reduced my queue request from 9 days to 3 days.

 

[StefanR5R]

It is okay if you beat me to 50M......no, really.

The latest trend is in your favor.
►graph

 

[Orange Kid]

Collatz Conjecture is the current sprint.

Less than a hour to the first Pentathlon project announcement.

 

[TennesseeTony]

Some information on Collatz is available here: http://teamanandtech.org/viewtopic.php?f=5&t=51

 

[ao_ika_red]

Some information on Collatz is available here: http://teamanandtech.org/viewtopic.php?f=5&t=51

re: crashtech reaction: You're a prophet! Will you show up in Crysis remastered? LOL

 

[TennesseeTony]

Collatz pays crazy points. I wonder how many points will separate the top 3 teams. Will it be tens of millions, hundreds of millions? A Billion point spread?

Combine that with bunkering and no lead is safe, lol. It is going to be rather entertaining, I think.

Place your bets, ladies and gentlemen.....I say first place and 3rd place will have a 3 Billion point gap at the conclusion.

 

[StefanR5R]

There was already a Collatz Conjecture sprint in 2017, but not in 2018-2019.
2017: GRC 579 M, CNT 520 M, TAAT 509 M, OCN 394 M, SUSA 250 M.
Of course, since then some individuals upgraded their GPUs.

 

[TennesseeTony]

SUSA has one guy who posted 450M points today. I think things have changed a bit since 2017.

 

[StefanR5R]

Sprint #5 at Collatz Conjecture — The first 18 hours.
League 1, top ten teams:

 

[StefanR5R]

PS,
million points per hour of teams 3...7:

 

[Orange Kid]

Collatz Conjecture day one.
The "Fasten Seatbelt" sign is lit . Looks to be a bumpy ride. Started 12th, then up to 2nd, then down to 5th, then up to 4th and now currently in 5th. Hang on all, who knows what is to come.



From FreeDC

 

[StefanR5R]

The second 18 hours.

 

[Orange Kid]

Collatz Conjecture day two plus.
Sorry for the late stats but was firing up the X6 out in the garage. Had to redo the whole dam thing as parts were remove and new parts added. Linux isn't as forgiving as winblows. All is well, so six more Rosetta cores and the ultimate goal of having pandora in the garage.
On with the stats..Back to 4th and cruising. Will we have enough to stay here, stay tuned!




From FreeDC

 

[StefanR5R]

The third 18 hours,
teams 1...11:


Hourly output of teams 3, 4, 5, 6, 8:

 

[biodoc]

Hmmm, somemone seems to have launched us into 2nd place. Nice drop @StefanR5R !

 

[StefanR5R]

The final 18 hours,
league 1, top 10 teams:


Teams 2...7, hourly output:

 

[biodoc]

Great looking graph! Nice job all!

 

[Orange Kid]

Collatz Conjecture final.
WOW, quite the ride. Second place!!! Fantastic finish all.



From FreeDC;
These stats seem to be lagging a bit and the real story isn't told here.

 

[biodoc]

These stats seem to be lagging a bit and the real story isn't told here.

Here's supplemental day 3 that captures some late bunker drops by @StefanR5R (272 M and change) , @Howdy and others.

 

[Orange Kid]

TN-Grid is the next sprint.