FormulaBOINC

[TennesseeTony]

Woo Hoo! Half an hour until we get to start BUNKERING for one day. The actual RACE won't start for another 24.5 hours.

I say, I say boy, it's got me in a singing mood!

Well, I came down there with my hat caved in
Oh I'll go back home with my pocket full of tin oh doo-dah day
Goin' to run all night, goin' to run all day
I'll bet my money on the bobtail nag, somebody bet on the bay.
Well, the Camptown ladies sing this song doodah doodah
Ah the Camptown race track's five miles long oh doo-dah day
Goin' to run all night, goin' to run all day
I'll bet my money on the bobtail nag, somebody bet on the bay.
Songwriters: Pat Irwin / Steven Foster
Camptown Races lyrics © Sony/ATV Music Publishing LLC

Sorry if you don't get the reference or the song. An old Saturday Morning cartoon character, Foghorn Leghorn, was well known for humming that old minstrel song about horse racing from the 1850s.

 

[TennesseeTony]

And the next project is: Rosetta@home

Be advised, on the web-preferences page, you can tell Rosetta@home how many hours you want your tasks to be. I think 8 hours is the default, changing that seems to have caused a lack of work for me in the past (changed it to 1 hour so I had better chances of finishing tasks before the race ended).

 

[Ken g6]

There's a "Target CPU run time" option for Rosetta. Mine is currently 12 hours, but I expect I'll want to lower that as the sprint gets close to the end.

 

[StefanR5R]

There is a long FAQ article about the "Target CPU run time" option, for those interested:
http://boinc.bakerlab.org/rosetta/forum_thread.php?id=669#10377

 

[TennesseeTony]

So the run time preference simply tells BOINC that you want to run the downloaded task at least once (for short times, like one or two hours), or to run it again and again to get different models out of the same task if you select a longer run time (better scientific results). Points are mostly time based, but they alluded that there was no set credit per task.

Also, they noted the problem we had with LHC (where the run time was from a few seconds to a day, thus confusing the BOINC client into timing out on legitimate tasks), and warn against changing your preference until you run out of tasks first, or you risk the long tasks getting the timed-out error. So don't change from one hour to 24 hours (just an example) without running out of 1 hour tasks first.

Further noted, the progress bar will 'get stuck' on 1%, then suddenly go to 100% and finish, if you run short duration tasks. This is not a problem, just something to be aware of, so don't panic and abort your tasks. The progress bar only changes after one model has been created, so if your time limit is set so short that the task only gets modeled once, it will go from 1%-100%. If set long enough to run the model 4 times, it will go from 1% to 25%, 50%, 75%, then 100%, as each model gets completed and checkpointed.

 

[StefanR5R]

During the half year in which I have been computing Rosetta on and off again, I haven't noticed jumps of percent completed. Maybe my runtimes were not short or long enough for this to occur, or they improved the progress reporting in the meantime. (FAQ article was last edited in 2006.)

 

[petrusbroder]

Added 28 cores to Rosetta and the TeAm.

 

[Orange Kid]

A quick look at where we are in the marathon stats

Rosetta@home...........Marathon.......Start
1 LinusTechTips_Team ...........25 ........16,469,437
2 Sicituradastra. .......................18 ........14,519,715
3 Overclock.net ........................15 ........12,422,299
4 TeAm AnandTech .................12 ........11,261,185
5 Team Norway ........................10 .........8,959,230
6 US NAVY ................................8 ..........8,602,426
7 France .....................................6 ..........6,691,648
8 Ars Technica ...........................4 ..........6,174,551
9 team.se ...................................2 ..........4,618,203
10 CANAL@Boinc .....................1 ..........4,590,747

 

[GLeeM]

Rosetta@home...........Marathon.......Start
1 LinusTechTips_Team ...........25 ........16,469,437
2 Sicituradastra. .......................18 ........14,519,715
3 Overclock.net ........................15 ........12,422,299
4 TeAm AnandTech .................12 ........11,261,185

There will be average points to make up, we have some but the others have more already running:
LTTT ............ 121,000
Sici ................ 81,000
OCN ............. 98,000
TeAm ............ 59,000

 

[Ken g6]

Well the first stats are up, and we're currently fourth.

http://formula-boinc.org/sprint.py?sprint=6&lang=en&year=2017

Looks like we have our work cut out for us. Anybody reading this is welcome to join the TeAm and help!

 

[crashtech]

I'm devoting what resources I can, though with the heat, I can't run things at 100%...

 

[iwajabitw]

I'm devoting what resources I can, though with the heat, I can't run things at 100%...

Family just arrived for the weekend. Had to shampoo the carpet and all, so I have the tsat set low. Brothers wife wants to know if its cooler than the inlaws at my place. I said nope, running the 280x's full steam with Rosetta, so its gonna be 78F at least. Glad its my weekend to work!!!!!!!

 

[GLeeM]

running the 280x's full steam with Rosetta

OK, Now I am confused - haven't run Rosetta since 2006 - I have a 280x that is a GPU.

Is there a CPU called 280x?

Or does Rosetta have GPU WUs?

 

[iwajabitw]

OK, Now I am confused - haven't run Rosetta since 2006 - I have a 280x that is a GPU.

Is there a CPU called 280x?

Or does Rosetta have GPU WUs?

No Rosetta gpu, but my dual cards are starving, so they are running with my Rosetta cpu tasks, since they take so little cpu time......Run Rosetta full CPU, sorry for the confusion, I wish they had a GPU client.

 

[crashtech]

Looks like OcUK is going to get the better of us unless we can dig up some more CPU power!

http://formula-boinc.org/sprint.py?sprint=6&lang=en&year=2017

 

[Ken g6]

Looks like OcUK is going to get the better of us unless we can dig up some more CPU power!

http://formula-boinc.org/sprint.py?sprint=6&lang=en&year=2017

The previous update they were 20K away. Now they're slightly over 10K away. I think we might pass them with the next update, at least temporarily.

 

[ocukguest]

Yes, it would seem you are about to overtake us..... although you never know whats around the corner

 

[Orange Kid]

After a slow start in fourth place we have worked our way slowly up to second. There a veiled threats coming from the OcUK camp though, to be careful as to what's around the corner. Hope their not in a dead end alley.

Rosetta@home.........Sprint.......Day One
1 Overclock.net...................... 25 ........601,903
2 TeAm AnandTech ...............18 ........428,937
3 OcUK - Overclockers UK .....15 .......407,906
4 Crunching@EVGA ...............12 .......355,228
5 LinusTechTips_Team ...........10 .......118,588
6 Sicituradastra. ........................8 ..........81,281
7 Team Norway .........................6 ..........64,976
8 US NAVY ................................4 ..........62,738
9 France .....................................2 .........52,170
10 meisterkuehler.de Team .......1 ..........47,974


Rosetta@home...........Marathon.......Start
1 LinusTechTips_Team ...........25 ........16,469,437
2 Sicituradastra. .......................18 ........14,519,715
3 Overclock.net ........................15 ........12,422,299
4 TeAm AnandTech .................12 ........11,261,185
5 Team Norway ........................10 .........8,959,230
6 US NAVY ................................8 ..........8,602,426
7 France .....................................6 ..........6,691,648
8 Ars Technica ...........................4 ..........6,174,551
9 team.se ...................................2 ..........4,618,203
10 CANAL@Boinc .....................1 ..........4,590,747

Rosetta@home...........Marathon.......Day One
1 LinusTechTips_Team ............25 ........16,704,088
2 Sicituradastra. ........................18 ........14,684,207
3 Overclock.net .........................15 .........13,145,563
4 TeAm AnandTech ..................12 .........11,738,272
5 Team Norway .........................10 ..........9,094,788
6 US NAVY .................................8 ...........8,726,150
7 France ......................................6 ...........6,796,413
8 Ars Technica ............................4 ...........6,267,208
9 team.se .....................................2 ...........4,681,776
10 CANAL@Boinc .......................1 ...........4,656,546

 

[StefanR5R]

Do you hear the battle noise from League 3? It's downright frightening what's going on over there. [H] take on the very top team of Rosetta@home.

 

[TennesseeTony]

I did notice League three is awake finally, and saw the Russians charge the hill and take down [H] earlier, but it's far from over! Way to go @[H]Coleslaw!


Regarding what's around the corner:
I once opened the kitchen door and stepped out into the garage, rounded the corner into the workshop, where a friggin' snake reared up like a Cobra and hissed at me! I screamed like a little girl and jumped back in surprise. Then I looked at the 8 inch long snake....and looked at my 12 inch long work boots attached to my feet....

 

[TennesseeTony]

Looks like I'm getting a ROUGH average of 90 points per 4 hour task.

 

[GLeeM]

Looks like I'm getting a ROUGH average of 90 points per 4 hour task.

I'm surprised that I am getting better than I have lately in:
DENIS@Home
LHC
TN-Grid
Cosmology
Stop
Accoustics

But not so good as CSGrid

 

[Orange Kid]

I'm getting 110'ish

 

[Skillz]

I did notice League three is awake finally, and saw the Russians charge the hill and take down [H] earlier, but it's far from over! Way to go @[H]Coleslaw!


Regarding what's around the corner:
I once opened the kitchen door and stepped out into the garage, rounded the corner into the workshop, where a friggin' snake reared up like a Cobra and hissed at me! I screamed like a little girl and jumped back in surprise. Then I looked at the 8 inch long snake....and looked at my 12 inch long work boots attached to my feet....

It's on like donkey kong! lol

 

[crashtech]

In other news, we've squeaked past meisterkuehler in Moo!, gaining 2 points:
http://formula-boinc.org/projets.py?lang=en&prj=145&year=2017

I could probably use a little help, this is a weird looking Pac-Man:
http://stats.free-dc.org/stats.php?page=team&proj=moo&team=173#map