Any interest in a SiDock challenge at BOINCStats?

[crashtech]

...Somehow I don't have high hopes in the responsibility particularly of the owner(s) of hosts with 2048 tasks in progress. We'll see.

Yeah I don't think so either. Users who abuse the system in such a way might be candidates for host banning imo.

@Fardringle , they know enough to use modified source code, with great power comes great responsibility.

 

[StefanR5R]

Originally, the SiDock server had the default setting of counting only up to 64 logical CPUs when determining the quota of tasks in progress. The admins lifted this at some point because there were a few contributors with 128- or 256-threaded computers.

(Edit)
BTW, if you over-bunkered but want to be a fair player and abort excess tasks, you really should abort and report them as early as possible. After you reported them, the project server might take quite a long time before assigning replicas to third hosts. AFAIK there are some arcane boinc server settings which influence this, and not all servers are set up specifically to give resends higher priority over tasks from newer workunits.

 

[cellarnoise]

while having only moderate or small CPUs (Core i5, i7, Ryzen…).

Hey! I resemble this! Not all of us can (or should I say do have?) have monster and efficient computers.... But what we lack in server hardware we also lack in production!

 

[StefanR5R]

I was thinking of 4-core i7 and 1st-gen Ryzen 3 or 5. :-)

But maybe it's just subconscious envy talking, given that I have only 'moderate or small' GPUs compared to many others here. :-D

 

[StefanR5R]

I am seeing a few Eprot tasks with names ending in _2 or _3 in my work buffers, intermixed with Eprot tasks called _0 and _1 and with 3CLpro tasks of course.

This indicates that a) at least some people took care to abort & report a bunch of their Eprot _0/1 tasks — for whatever reason, e.g. work buffers set too large. And b) the project server created replicas and mixed them among tasks from new Eprot and 3CLpro workunits in a timely manner — which is good.

PS, some users might bunker in a way that results are not reported but already uploaded. Such users would be hesitant to report any aborted tasks early. They couldn't report aborted tasks separately from completed results (without client source code modification which would go a bit further than finding one or another numeric constant in the code).

 

[Kiska]

And b) the project server created replicas and mixed them among tasks from new Eprot and 3CLpro workunits in a timely manner — which is good.

Yes that is good since it means the transitioner is performing as expected and expediently

Oh btw with the runtime graph seconds or hours?


The server reports the figure as hours, I just convert it to seconds. But how would you like to see the number represented?(as hours or keep it as seconds)

 

[StefanR5R]

But how would you like to see the number represented?(as hours or keep it as seconds)

I'm not sure.

There was once an effort to decimalize the time units, but it went nowhere. :-|

 

[cellarnoise]

This SiDock challenge and banter is so boring that I will in the future participate more on the expected interaction than the points or whatever.

Sad.

If this is what the "Singularity" is about, the puters will all shut down by self choice!

 

[StefanR5R]

if you over-bunkered but want to be a fair player and abort excess tasks, you really should abort and report them as early as possible. After you reported them, the project server might take quite a long time before assigning replicas to third hosts. AFAIK there are some arcane boinc server settings which influence this, and not all servers are set up specifically to give resends higher priority over tasks from newer workunits.

PS, about that last part: I recall that a server might be set up to assign resent tasks to "reliable" hosts only. QuChemPedIA once had a period during which the server's criteria were strict enough, and the error rate of their application high enough, that there were no longer any hosts considered "reliable". And so the tasks which were to be resent lingered on the server for weeks, or months even.

This SiDock challenge and banter is so boring

To compensate and cater at least a tiny bit to the excitement, the BoincStats site is not trying too hard to pull stats, and now and then sets some team points (and all user points) to zero instead. >:-)

 

[lane42]

6
Days13
Hours15
Minutes59
SecondsSince start of challenge
10
Hours44
Minutes1
SecondTill end of challenge

Team Name	Rank	Credit	-1:00	-2:00	-4:00	-8:00	-16:00	-32:00	Options
TeAm AnandTech	1	9,140,293	9,055,683	8,986,481	8,846,809	8,522,432	8,002,861	7,082,845
[H]ard|OCP	2	6,278,951	6,216,960	6,174,002	6,090,888	5,873,257	5,493,577	4,801,829
Ukraine	3	2,179,709	2,165,227	2,151,618	2,127,109	2,059,908	1,975,645	1,799,229
BOINC.Italy	4	2,010,651	2,000,187	1,993,212	1,977,308	1,937,858	1,876,450	1,715,739
BOINC@Poland	5	1,330,196	1,320,754	1,314,449	1,300,612	1,258,724	1,194,487	1,073,665
Crystal Dream	6	1,210,634	1,197,795	1,189,461	1,176,441	1,145,034	1,074,887	955,145
AMD Users	7	760,842	753,057	747,805	734,786	707,566	655,650	578,684
BOINC@AUSTRALIA	8	726,040	719,989	713,306	704,294	677,976	633,166	541,895
UK BOINC Team	9	367,747	364,261	361,904	355,383	342,386	319,825	271,016
BOINCstats	10	321,557	318,070	315,528	309,813	297,962	276,045	240,866

Something must be a little off.

 

[StefanR5R]

Something must be a little off.

The SiDock site is sometimes laggy today (users begin to remember that they wanted to report some results), and BoincStats doesn't tolerate lag very well. It has corrected itself after your post though... until the next glitch.

 

[Fardringle]

The SiDock site is sometimes laggy today (users begin to remember that they wanted to report some results), and BoincStats doesn't tolerate lag very well. It has corrected itself after your post though... until the next glitch.

Yep. For a good part of the day on Sunday it was showing that we and several other teams had zero total credits...

 

[Fardringle]

With exactly 1.5 hours left, we have a 1.1M lead over SETI.Germany. Still time for someone to drop a huge bunker, but maybe not time for it to be validated and posted to the stats...

Team Name	Rank	Credit	-1:00	-2:00	-4:00	-8:00	-16:00	-32:00	Options
Planet 3DNow!	1	17,766,892	17,687,268	17,590,350	17,350,701	16,937,141	-	14,686,129
TeAm AnandTech	2	9,865,302	9,812,534	9,736,785	9,531,896	9,232,304	-	7,587,480
SETI.Germany	3	8,759,160	8,727,980	8,633,489	8,177,725	7,575,179	-	6,274,600
Rechenkraft.net	4	6,757,993	6,721,023	6,684,705	6,590,195	6,416,049	-	5,457,824
[H]ard|OCP	5	6,693,040	6,664,582	6,624,929	6,528,431	6,342,308	-	5,177,404
SETI.USA	6	5,124,479	5,098,673	5,068,267	4,996,390	4,853,406	-	4,067,256
The Scottish Boinc Team	7	3,661,167	3,636,295	3,609,233	3,528,440	3,381,878	-	2,182,336
L'Alliance Francophone	8	3,318,181	3,303,616	3,284,048	3,232,514	3,147,302	-	2,699,608
Ukraine	9	2,321,833	2,307,670	2,293,377	2,260,277	2,195,019	-	1,897,320
BOINC.Italy	10	2,087,082	2,082,009	2,075,503	2,060,024	2,022,836	-	1,810,893

 

[Fardringle]

Final stats. Nice job, team!

Team Name	Rank	Credit	-1:00	-2:00	-4:00	-8:00	-16:00	-32:00	Options

Planet 3DNow!	1	17,904,271	17,827,040	17,747,095	17,564,238	17,161,015	16,365,333	-
TeAm AnandTech	2	9,993,841	9,913,808	9,850,479	-	9,399,491	8,790,684	7,675,635
SETI.Germany	3	8,816,140	8,785,397	8,750,716	-	8,046,335	7,211,760	6,327,288
Rechenkraft.net	4	6,810,163	6,783,567	6,747,047	-	6,511,410	6,164,185	5,520,767
[H]ard|OCP	5	6,772,559	6,728,473	6,684,239	-	6,442,381	6,050,866	-
SETI.USA	6	5,197,210	5,160,342	5,116,061	-	4,937,226	4,633,763	-
The Scottish Boinc Team	7	3,714,931	3,689,934	3,654,269	3,603,728	3,449,362	3,110,636	2,242,684
L'Alliance Francophone	8	3,349,537	3,332,050	3,314,656	-	3,193,921	3,038,218	-
Ukraine	9	2,340,735	2,331,765	2,317,831	-	2,226,928	2,111,280	-
BOINC.Italy	10	2,096,724	2,091,392	2,085,968	-	2,045,632	1,969,943	-

 

[crashtech]

I'm glad I did not pull all the stops, I think P3D wanted this one pretty badly.

 

[cellarnoise]

Probably the least amount of entertainment I have had at a D.C. event. Though I'm only like 2 years in?

With the late notice, screwed up start and poor work server for the event dishing work, and not much as far as easily advertised event stats or other stats...

Sorry I am bitching, but even within the AnandTeam, not as much fun as I am have liked in the past. I like more chatter and I will now try to be more reserved, though that may be liked... Though I bitch plenty and with good muscle relaxers can get a bit crazy.

Too bad my 8 billion point bunker won't get registered in time. I was hoping to get some "block choy" or "parsnip" soup, but it just seems too much... My supercharged and overblocked 1700x cpu work will go to my waist in the form of beer. Haha and kidding.

 

[pututu]

Congratz on finishing 2nd.

 

[lane42]

Probably the least amount of entertainment I have had at a D.C. event. Though I'm only like 2 years in?

Well, Sidock isn't the most exciting project around

 

[StefanR5R]

We received our badge now:



(team page)


Subproject status:
............................................................ Sep 16 ........ Sep 23
Target: corona_Eprot_v1 (%) ............ 60.612 ........ 65.214
Target: corona_3CLpro_v4 (%) ......... 40.037 ........ 50.723
Target: corona_3CLpro_v5 (%) ......... 39.807 ........ 50.692
Target: corona_3CLpro_v6 (%) ......... 39.626 ........ 50.400

 

[cellarnoise]

That is one SCARY smiley face badge!

Was this a tribute to Stephen King challenge? Sorry, I have not been tracking as there was very little team banter.

 

[StefanR5R]

Fooled ya! >:-D

 

[StefanR5R]

There is another competition held by SiDock again:
— announcement — contest page —

The best three teams will win unique badges again.

The timing:

• Halloween — HCHC — details t.b.a. about 7 hours from now

• 11/02 – 11/09 (11:45 UTC) SiDock November competition

• about a week before 11/16: bunkering for WCG B'day
• 11/16 – 11/23 (00:00 UTC) World Community Grid Birthday Challenge

• 11/23 - 11/26 (05:00 UTC) PrimeGrid AP27 challenge

Practically it fits, but would make for a tight schedule.

So to revisit the thread subject:
Any interest in a SiDock challenge at BOINCStats?

————————
PS:
There are already 12 teams subscribed to this challenge, even though it was just announced. But e.g. in case of SETI.Germany, they are subscribed because they have an automatism in place which subscribes them to each Boincstats challenge. I.e., at least in SG's case, their subscription is not an indicator for their interest. I don't know if more teams auto-subscribe to Boincstats challenges.

 

[lane42]

SiDock November competition

My small fleet says yes!!

 

[Skivelitis2]

Sounds good, I'll be in. Hopefully they stick to the short work units. Those monsters during the Pentathlon were rough on my old hardware.

 

[Fardringle]

I'll join with my limited resources.