Formula Boinc Sprints 2018

[Kiska]

Actually I stand corrected, I don't think the work generator is running.
This is the last WU I was able to lookup without it going "unable to handle request"
http://asteroidsathome.net/boinc/workunit.php?wuid=88871265
The unit was created on the 15th of June

Also by using the method of lookup I did, I am able to infer the work generator is doing 25 WU per second, resulting in a task generation of 50 per second when its on

 

[StefanR5R]

Disclaimer: I like your post, not how the server is operated.* :-(

The two tasks of this WU were sent 6h20m after the sprint project was announced. This coincides with the time I saw mentions that the project no longer has tasks in European forums. (It was in the morning in Europe and in the middle of the night in America.)

Since we began receiving tasks again, which was shortly after the beginning of the sprint, we have been given exclusively 3rd/ 4th/... tasks of WUs which had prior tasks returned with error, canceled by the client, or timed out.

*) Edit: It's completely OK if a project issues work only intermittently. But when they do, they shouldn't be selected for a sprint, obviously.

 

[Kiska]

Since we began receiving tasks again, which was shortly after the beginning of the sprint, we have been given exclusively 3rd/ 4th/... tasks of WUs which had prior tasks returned with error, canceled by the client, or timed out.

These tasks are described as ps_(input)_input_(segment)_(segment)_(BOINC task number)

so eg if that WU had a timed out task it would be described as: ps_180614_input_99_9_2 or ps_180614_input_99_9_3, etc etc etc

 

[Kiska]

Oh look, the server is generating work!

 

[Kiska]

OK I have taken 3 samples of generation rate

From task IDs: 208283670 - 208283701 these IDs are from 16:01:41 UTC, and I observed 31 tasks generated per second
IDs: 208283702 - 208283741 from 16:01:42 UTC, observed 39 tasks per second
IDs: 208283742 - 208283775 from 16:01:43 UTC, observation: 33 tasks per second

From this limited sample, task generation is 34.3 tasks per second

 

[Kiska]

Also if you wanna observe a WU without tasks here is that URL: http://asteroidsathome.net/boinc/workunit.php?wuid=89105947

It will take about 1.5 hrs before the task generator will get to this unit. From 17:14:14 UTC

 

[StefanR5R]

A fairly boring day for us, but check out the UK vs. Czech battle!

I do wonder though whether CNT aren't doing only the bare necessary here in league 1, and are meanwhile guest-crunching for their kinsmen in league 3 in the vigorous battle against the pirates that is raging over there.

 

[StefanR5R]

The last 18 hours:

Forgive me if I call our distance from the other teams (in this league) astronomical.

 

[Orange Kid]

Final for Asteroids@home
Well that's AMAZING !!!! A second place.
Lets hope the next one is a little better managed by the project picked.
Many thanks to all. A fantastic TeAm we have here

1 Gridcoin .........…............…….25 ............25,715,040
2 TeAm AnandTech .........…….18 ............12,582,720
3 SETI.USA ..............................15 .........….8,477,280
4 Czech National Team ............12 .........….4,031,520
5 UK BOINC Team ...............….10 .........….3,985,920
6 Rechenkraft.net ........................8 .........….1,993,920
7 SETI.Germany ............…......….6 .........….1,802,400
8 Planet 3DNow! .....................….4 .........….1,660,800
9 L'Alliance Francophone ............2 .........….1,135,680
10 SETIKAH@KOREA ............….1 ............….994,080

The F1 results
June 24: French Grand Prix
1. Lewis Hamilton, Mercedes
2. Max Verstappen, Red Bull
3. Kimi Raikkonen, Ferrari

 

[Ken g6]

It's completely OK if a project issues work only intermittently. But when they do, they shouldn't be selected for a sprint, obviously.

I dunno, I might not mind Xansons being selected for a sprint. Just not this week, obviously. 10 minutes or so until the next announcement!

 

[biodoc]

The Sprint project is Rosetta@home

 

[Markfw]

The Sprint project is Rosetta@home

What day/time does this start ? I am already in for over 300 cores for Rosetta, just curious.

 

[Ken g6]

@Markfw, see if you can suspend networking until this time tomorrow.

 

[TennesseeTony]

Sheesh, Rosetta is a data cap buster in case anyone has limited data....

 

[StefanR5R]

What Ken said; or more roundabout:
from Thursday, 06/28/2018 22:00 (UTC) to Sunday, 07/01/2018 21:59 (UTC) =
from Thursday, 06/28/2018 15:00 (PDT) to Sunday, 07/01/2018 14:59 (PDT),
if I am not mistaken.

The sprint begins after the following table was filled with the team names:
http://formula-boinc.org/sprint.py?sprint=9&year=2018

 

[Ken g6]

Sheesh, Rosetta is a data cap buster in case anyone has limited data....

If you hurry while the first couple of files are downloading you can copy a bunch of the other files from another machine.

 

[crashtech]

Unfortunately, most of my running TN-Grid tasks would expire before the end of the Sprint, so I feel that they must be allowed to complete. This will slow me down a bit.

 

[Markfw]

@Markfw, see if you can suspend networking until this time tomorrow.

Done. Please remind me tomorrow at the time to un-suspend. I have my BCG treatment tomorrow, and I become a vegetable.
(my inert tuberculosis cancer treatment)

 

[StefanR5R]

Outcome of the Rosetta@home sprint in May 2017:

Gridcoin .................................. 3.85 M
SETI.USA .................................. 3.41 M
Czech National Team ............. 2.94 M
TeAm AnandTech ................... 2.04 M
L'Alliance Francophone ......... 1.82 M
Overclock.net .......................... 1.78 M
OcUK - Overclockers UK ........ 1.36 M
USA .......................................... 0.88 M
UK BOINC Team ..................... 0.86 M
BOINC@Poland ...................... 0.69 M

(Notable 2017 sprint results of other teams who are not in 2018's league 1:
TSC!Russia 2.19 M, [H]ard|OCP 2.08 M, Crunching@EVGA 1.03 M, XtremeSystems 0.0005 M.)

 

[crashtech]

I had to pull my oldest Ryzen, a 1700X, off of Rosetta due to their being more computation errors than anything else. It may be that I need to back down the memory timings, but it's headless, so it will be a while before that happens.

Links to the logs of some of the bad WUs:
https://boinc.bakerlab.org/result.php?resultid=1010290990
https://boinc.bakerlab.org/result.php?resultid=1010291000
https://boinc.bakerlab.org/result.php?resultid=1010291045
https://boinc.bakerlab.org/result.php?resultid=1010290809
https://boinc.bakerlab.org/result.php?resultid=1010290810
https://boinc.bakerlab.org/result.php?resultid=1010290813

 

[Ken g6]

Done. Please remind me tomorrow at the time to un-suspend.

@Markfw, it's now safe to upload your results. Release the hounds! (Any time you're ready.)

 

[Ken g6]

And we're off to a decent start, in second place, behind Gridcoin.

 

[Markfw]

And we're off to a decent start, in second place, behind Gridcoin.

Where do you see the points ? wondering what my bunker was worth

On this page: http://formula-boinc.org/sprint.py?sprint=9&year=2018

We are first !

 

[TennesseeTony]

So nice of the other teams to be asleep for the most part, still clutching their bunkers, while we very temporarily get to be in first place.

 

[Markfw]

well, that didn't last long. At least we are in second.