2020 Formula Boinc Sprints

[Endgame124]

I think all of my hosts have tasks again. The big producers never ran out, but I don't let my work queue get very long on the raspberry pis, so I had 6 that didn't have any rosetta jobs. Hopefully that will bump my production up a little more.

 

[Assimilator1]

Good to see I'm in there , not sure I've been much help though as my rig ran out of WUs, and when I saw they had no WUs I allowed new tasks for LHC, I've since turned them off again but it's been finishing off the remaining tasks. I'll suspend LHC in a few minutes to allow a WU to finish.

 

[Orange Kid]

Rosetta@home day two.

1	Gridcoin	25	4,565,876
2	TeAm AnandTech	18	4,313,779
3	XtremeSystems	15	3,625,543
4	USA	12	3,193,284
5	Dutch Power Cows	10	2,105,316
6	UK BOINC Team	8	1,732,830
7	The Scottish Boinc Team	6	1,254,208
8	Rechenkraft.net	4	1,067,714
9	Czech National Team	2	901,172
10	Team 2ch	1	879,153

FreeDC not yet updated.
EDIT: FreeDC update

Mark_F_Williams	523,248	396,485		919,733
user_2072655	794,227	4		794,231
xii5ku	433,227	341,137		774,364
crashtech	143,022	216,636		359,658
10esseetony	283,984	74,057		358,041
biodoc	138,906	182,031		320,937
phoenicis	67,486	60,556		128,042
Howdy	85,384	0		85,384
boosthacker	45,317	25,927		71,244
Orange Kid	25,723	24,654		50,377
Endgame124	23,920	21,602		45,522
far	33,199	4,627		37,826
Closius	19,753	12,637		32,390
BadThad	15,775	16,437		32,212
Fardringle	28,075	0		28,075
[H]Skillz	15,234	9,575		24,809
Ken_g6	20,785	3,781		24,566
Bullseye	18,658	4,063		22,721
Rudy Toody	19,381	0		19,381
kart17	7,355	7,865		15,220
voodoo5_6k	11,874	2,984		14,858
Lane42	10,517	0		10,517
Wiz	4,802	4,866		9,668
TeAm Enterprise	4,584	4,146		8,730
Mike	5,313	2,904		8,217
Bremen	4,534	2,268		6,802
mnelsonx	6,232	0		6,232
uallas5	4,735	511		5,246
HappyCracker	2,113	2,676		4,789
cyost91	3,448	887		4,335
geoffp	4,171	0		4,171
Quoll	478	2,466		2,944
blckgrffn	2,625	0		2,625
Cassio Hui	1,863	0		1,863
Viperoni	869	977		1,846
Dave	946	721		1,667
TA_JC	1,498	0		1,498
geecee	401	989		1,390
kenneth	808	251		1,059
Mort	206	475		681
KennedyIO	248	0		248

 

[Markfw]

who is user_2072655 ? I thought that with over 500k points, I would be the number one contributor for Rosetta, but this guy just blew my pants off...With 800k for the day....

I mean I have 800 cores/threads, this guy must have 1200 or more.

 

[VietOZ]

who is user_2072655 ? I thought that with over 500k points, I would be the number one contributor for Rosetta, but this guy just blew my pants off...With 800k for the day....

I mean I have 800 cores/threads, this guy must have 1200 or more.

nah, just a couple days bunker. You're still king on Rosetta
and it wasn't me btw lolz

how's everyone doing? Stay safe my fellow crunchers! It's crazy out there

 

[StefanR5R]

Rosetta's work availability:

(source: Kiska)

Gridcoin's daily Rosetta output:

(source: BOINCstats)

You can spot the resemblance.
Their output should be back up at about 3.5 M PPD shortly.

 

[TennesseeTony]

who is user_2072655 ? I thought that with over 500k points, I would be the number one contributor for Rosetta, but this guy just blew my pants off...With 800k for the day....

I mean I have 800 cores/threads, this guy must have 1200 or more.

We don't know who it is, but I believe it is the same user that joined during the Pentathlon. And yes, they are one heck of a powerhouse!


Thanks for the stats, Orange Kid!

 

[Assimilator1]

I seem to be missing from the stats today , despite switching back to Rosetta last night........

 

[lane42]

I seem to be missing from the stats today

Are you user_2072655 J/K

 

[Markfw]

Are you user_2072655 J/K

I hope that user is bunkering for us, and dumps 800k in the next few hours like last time.

 

[Endgame124]

How do you manage to bunker for Rosetta? I perpetually have tasks in my queue with less than 24 hours until the deadline. Do you stop accepting new tasks, wait until 3 days before event start and then start accepting tasks again?

 

[StefanR5R]

I saw only tasks with 3 days deadline myself so far. Are these perhaps resends, i.e. replicas from WUs which received errors from other computers? — In general, when you prepare a bunker it is always good to remember to check the queue for tasks with less than the desired deadline, abort & report them, and request more work if your queue has gotten too short due to this. It is also possible to abort & report while network transfers are suspended in the client (you can't request replacement work then, of course).

 

[Endgame124]

I saw only tasks with 3 days deadline myself so far. Are these perhaps resends, i.e. replicas from WUs which received errors from other computers? — In general, when you prepare a bunker it is always good to remember to check the queue for tasks with less than the desired deadline, abort & report them, and request more work if your queue has gotten too short due to this. It is also possible to abort & report while network transfers are suspended in the client (you can't request replacement work then, of course).

I’m always running Rosetta, and my queues are set such that I keep about 1.5 extra days of work on hand. Generally that means I’m submitting work units with 30 or less hours until the deadline. If, 3 days before the start of an event I increase to 3 days of work on and turn off networking, I’ll end up with a bunch of tasks that fails.

Sounds like, in that case, people are aborting those tasks to free up space for fresh ones with a deadline just after event start?

 

[StefanR5R]

Only if they are aware of that there may be tasks with less than the usual deadline, and remember to check while building the workqueue for the bunker.

BTW, some people might want to bunker until shortly after the start of a contest ... and other people might want to bunker until shortly before the end of a contest.

I am having client instances active in this Rosetta sprint which haven't been running Rosetta for a long while, or were even entirely new to Rosetta. (The primary client instances of my bigger computers are keeping suspended QuChemPedIA tasks which were interrupted by this sprint.) The Rosetta@home scheduler certainly sends such special high-priority tasks with reduced deadline only to hosts which were constantly active recently and had short turnaround periods. And therefore none of my clients qualified for such tasks and I saw none, but your hosts qualify.

 

[Orange Kid]

Rosette@Home Final
Great sprint all and a valiant effort, but the GC army is formidable.

1	Gridcoin	25	8,087,640
2	TeAm AnandTech	18	7,532,426
3	XtremeSystems	15	5,601,477
4	USA	12	4,998,466
5	Dutch Power Cows	10	3,521,177
6	UK BOINC Team	8	3,186,389
7	The Scottish Boinc Team	6	2,029,283
8	Rechenkraft.net	4	1,834,087
9	Team 2ch	2	1,593,922
10	Czech National Team	1	1,574,748

FreeDC has not yet updated.

and during my sleepy time;

user_2072655	829,980	794,227			1,624,207​
Mark_F_Williams	685,196	523,248	396,485		1,604,929​
xii5ku	476,479	433,227	341,137		1,250,843​
10esseetony	317,533	283,984	74,057		675,574​
crashtech	163,390	143,022	216,636		523,048​
biodoc	131,050	138,906	182,031		451,987​
Howdy	119,639	85,384			205,023​
phoenicis	75,602	67,486	60,556		203,644​
boosthacker	33,197	45,317	25,927		104,441​
Orange Kid	40,681	25,723	24,654		91,058​
far	40,593	33,199	4,627		78,419​
Endgame124	26,129	23,920	21,602		71,651​
Ken_g6	42,823	20,785	3,781		67,389​
Fardringle	36,643	28,075			64,718​
Icecold-Team Anandtech	59,301	0			59,301​
Closius	17,543	19,753	12,637		49,933​
Rudy Toody	27,547	19,381			46,928​
Bullseye	21,740	18,658	4,063		44,461​
BadThad	12,013	15,775	16,437		44,225​
[H]Skillz	16,579	15,234	9,575		41,388​
voodoo5_6k	12,974	11,874	2,984		27,832​
Lane42	16,104	10,517			26,621​
kart17	7,676	7,355	7,865		22,896​
[TA]Assimilator1	11,195	0	6,884		18,079​
ZipSpeed	17,713	0			17,713​
mnelsonx	10,185	6,232			16,417​
Mike	8,102	5,313	2,904		16,319​
Wiz	4,375	4,802	4,866		14,043​
TeAm Enterprise	4,499	4,584	4,146		13,229​
cyost91	7,876	3,448	887		12,211​
geoffp	7,681	4,171			11,852​
EdvardasR	11,584	0			11,584​
Bremen	2,293	4,534	2,268		9,095​
uallas5	3,493	4,735	511		8,739​
HappyCracker	3,473	2,113	2,676		8,262​
blckgrffn	4,659	2,625			7,284​
Cassio Hui	1,806	1,863			3,669​
TA_JC	2,150	1,498			3,648​
Bundyman	3,542	0			3,542​
Quoll	0	478	2,466		2,944​
Viperoni	982	869	977		2,828​
Dave	1,045	946	721		2,712​
geecee	1,270	401	989		2,660​
kenneth	1,246	808	251		2,305​
John	1,992				1,992​
fidgety	1,402	0			1,402​
Mort	129	206	475		810​
iotone	504				504​
Jesper S. Rytke	386	0			386​
KennedyIO	0	248			248​

 

[lane42]

Thanks for the stats OK
Great going Team

 

[StefanR5R]

We don't know who it is, but I believe it is the same user that joined during the Pentathlon. And yes, they are one heck of a powerhouse!

BTW, the user was already with us before Pentathlon, and helped us during the entire duration of Pentathlon 2020's Rosetta@home Marathon. (Those who are really curious can find this user's Rosetta@home performance during Pentathlon in boincstats' table of "Last 60 days", and their performance at the 11/2019 Formula BOINC sprint at Rosetta@home in FreeDC's table of "recent milestones".)

On another note,
here are our results at Rosetta@home sprints at Formula BOINC, complete since the introduction of the sprint format:

05/2017 ........ 2.0 M ........ Gold in League 2
06/2018 ........ 3.4 M ........ Bronze in League 1
11/2019 ........ 7.7 M ........ Silver in League 1
06/2020 ........ 7.5 M ........ Silver in League 1​

I think it's fair to say that TeAm AnandTech's performance in Rosetta@home contests has been pretty good recently. :-)

Our 2020 result was of course somewhat impacted by work scarcity during ≈2 of the 1+3 days. But Gridcoin for example dropped from 10 M in 2019 to 8 M in 2020. While this may have various causes, their output before and after the shortage points to the task shortage to be the major or only reason for their reduced performance this year.

Oh, compared with the 1+3-day sprint results above, our 29 M at the Rosetta@home Marathon of Pentathlon 2020 (3+14 days) are rather excellent. That was 1.7 or 2.1 M/day average, depending on whether you divide this by 17 or 14 days.

 

[Ken g6]

Looks like some Einstein picked another GPU project for the summer.

Yes, that project is Einstein@Home.

 

[Orange Kid]

Einstein@Home is the next sprint.

 

[TennesseeTony]

We came in 2nd behind XS last year, SUSA was 3rd. I think the scores were 120M, 102M, and 93M. Pretty close challenge.

My budget isn't going to allow a full fleet GPU race, but I can run the 2080Ti.

 

[StefanR5R]

Looks like some Einstein picked another GPU project for the summer.

In the morning, before hurrying to work, I got mere ≈180 W wrung out of my 250 W cards. I took care to configure the computers for 100 % shader utilization though.

I left the windows open, for bugs and birds to come and take a peek.

We came in 2nd behind XS last year, SUSA was 3rd. I think the scores were 120M, 102M, and 93M. Pretty close challenge.

UBT didn't make it into the top ten of the 2019 Einstein sprint, but might make it this year. ;-)

 

[biodoc]

I left the windows open, for bugs and birds to come and take a peek.

 

[Assimilator1]

I left the windows open, for bugs and birds to come and take a peek.

Lol, I hope you left them some snacks too!

Oh, I'm in on Einstein, crunching now.
[edit] And just added my 2nd rig to it too, seeing as it's something it's good at, give that old Tahiti a work out! (so tempted to get the HD 7970 out of it's packaging! But I've got it up for sale atm).

 

[Ken g6]

I left the windows open, for bugs and birds to come and take a peek.

Which goes to show, GPU projects in the summer are for the birds!

I keep getting errors on Gravitational Wave search WUs. Don't know what's going on. My GPU is off until I can dump my results and reset the project.

 

[Endgame124]

I just set one of my hosts up, but figured I would let the active folding wu wrap up first.