[TSC] hitparade van 10 november Vorige deel Overzicht Volgende deel Laatste deel

Dit topic is onderdeel van een reeks. Ga naar het meest recente topic in deze reeks.

Pagina: 1
Acties:

Onderwerpen


Acties:
  • 0 Henk 'm!

  • MarkB
  • Registratie: Juni 2000
  • Niet online
Een zeer mooie flush van TCF. Dat moest ook wel anders was onze score helemaal dramatisch geweest. Controleren jullie of deze scores goed zijn dan verzin ik achteraf wel een verhaaltje :+

TSC hitparade van 10 november 2003

Top 30

1.(+)3.192DPC_Team_ColdFusion118.227(2)
2.(+)703DPC_Noizy_Cows135.771(1)
3.(+)470DPC_Kippie114.992(3)
4.(+)449Elteor_Borislavj31.112(9)
5.(+)332We_R_cow14.373(14)
6.(+)178glurp_oc70.323(6)
7.(+)112derrie35.220(7)
8.(+)99bepaalde_iemanden71.499(5)
9.(+)26pool692.887(30)
10.(+)18eaglebeast4.515(25)
11.(+)17MsentinelB3.670(27)

Overall top 30

1.(=)135.771DPC_Noizy_Cows703(2)
2.(=)118.227DPC_Team_ColdFusion3.192(1)
3.(=)114.992DPC_Kippie470(3)
4.(=)72.361DPC_Team_Switch0(0)
5.(=)71.499bepaalde_iemanden99(8)
6.(=)70.323glurp_oc178(6)
7.(=)35.220derrie112(7)
8.(=)31.419DPCquatar0(0)
9.(=)31.112Elteor_Borislavj449(4)
10.(=)30.408NC_TigZ0(0)
11.(=)29.585kuhkuh0(0)
12.(=)17.360DPC_Kinslayer0(0)
13.(=)16.713QUCCC0(0)
14.(=)14.373We_R_cow332(5)
15.(=)13.346DPC_Team_Westfriesland0(0)
16.(=)10.068ArmEagle0(0)
17.(=)8.935LPC_Je_oen0(0)
18.(=)8.374TheFly0(0)
19.(=)7.339DPC_Clan_of_ToeT0(0)
20.(=)6.382VFRmaniac0(0)
21.(=)6.008Delton0(0)
22.(=)5.334Scorpion0(0)
23.(=)5.327DPC_Maniac_Team0(0)
24.(=)4.854DPC_Team_Aggression0(0)
25.(=)4.515eaglebeast18(10)
26.(=)4.330Ruffryders0(0)
27.(=)3.670MsentinelB17(11)
28.(=)3.541DPC_dynamite0(0)
29.(=)3.500Heineken0(0)
30.(+1)2.887pool6926(9)

Top 15 teams

1.(+4)138.488UDGermany @ TSC1.682.134(4)
2.(-1)15.598Russia3.602.323(1)
3.(-1)13.606BroadbandReports.com...2.655.959(2)
4.(-1)6.738TSC@Hope2.546.569(3)
5.(-1)6.229Bodensee-Team1.568.667(5)
6.(+4)6.087TSDeV-Koeln-Duesseld...438.114(10)
7.(+1)5.596Dutch Power Cows925.871(8)
8.(-2)5.225The Possums1.337.314(6)
9.(+74)3.190Team Leipzig10.774(72)
10.(+28)922Griffin's Gang42.889(37)
11.(+9)879AMD Users115.566(20)
12.(=)839F. Karilon - Italia316.264(12)
13.(+10)784Lithuania104.663(22)
14.(+43)739Albert Schweitzer17.226(56)
15.(+11)696Ars Technica Team Fr...76.853(26)

Leden

Aantal leden102
Verschil met vorige dag0
Nieuwe leden0
Retirements0

Huh?? Niemand joinde gisteren DPC!!! Boehoehoehoeeeee.... :'(

De oude crew... helemaal zo slecht niet.


Acties:
  • 0 Henk 'm!

  • ColdFusion
  • Registratie: Augustus 2000
  • Laatst online: 30-05 12:47
De scores kloppen op zich wel, namelijk gewoon de scores van gisteren. Er mist alleen de produktie van een dag minimaal (zondag). Het lijkt erop alsof wel weer de results zijn aangekomen (dus niet verloren), maar dat ze niet kunnen vertellen van wie ze waren, dus geen stats daarvan. Het is nog even volhouden met dit oude statssysteem bij TSC (en D2OL). De ontwikkeling van een nieuw statssysteem staat geplanned voor Q1 2004.

_______________-=Team ColdFusion=-_______________
#2 DPAD - #2 OGR27 - #2 F@H - #3 R@H - #5 RC5 - #6 SoB - #10 WCG
||| #1 ECCp109 - #1 ECC2-109 - #1 OGR24 - #1 D2OL P1|||


Acties:
  • 0 Henk 'm!

  • mcB
  • Registratie: Mei 2002
  • Laatst online: 11:20

mcB

_/-\o_ MarkB voor de TSC stats en TCF voor de heerlijke flush
We 'R cow flusht door, maar het moet echt harder,... met ons en met DPC!

Strix (Skullflame)


Acties:
  • 0 Henk 'm!

  • SlinkingAnt
  • Registratie: December 2001
  • Niet online
Hmm, door dat grappie ben ik wel minstens 200 candidates armer :(

Intel C2Q 9450@3.3 | Gigabyte P35-DS4 | Sapphire R280x | 4x 2GiB PC6400 Kingston DDR2 | 1x Intel 320SSD 240GB | 2x Spinpoint F1 320GiB


Acties:
  • 0 Henk 'm!

  • Flapmo
  • Registratie: April 2000
  • Laatst online: 20:13

Flapmo

and back is gigi!

TCF komt eraan :(.

"The purpose of computing is insight, not numbers." -- Richard Hamming


Acties:
  • 0 Henk 'm!

  • Bastien
  • Registratie: Augustus 2001
  • Niet online

Bastien

Probleemeigenaar

Het duurt even, maar dan heb je ook wat :P

Mja, als het zo moet met die candidates... flush ik maar niet meer van vrijdag tot maandag... raar verhaal imo :/

Zolang er mensen zijn zal de wereld nooit duurzaam worden.
Zijn er wel genoeg kliffen op de wereld waar 8 miljard mensen zich als lemmingen vanaf kunnen storten?


Acties:
  • 0 Henk 'm!

  • mcB
  • Registratie: Mei 2002
  • Laatst online: 11:20

mcB

Ach zulke probs met de stats overkomt de besten.
Zie Dnet,... die hadden/hebben de zaakjes goed voor elkaar. Zelfs daar zijn er de afgelopen tijd regelmatig problemen geweest met de stats.

En als in het eerst kwartaal van het komend jaar aan de TSC stats gewerkt gaat worden, kunnen ze terugkijken op problemen in het verleden en er lering uit trekken.

Strix (Skullflame)


Acties:
  • 0 Henk 'm!

  • JustMe
  • Registratie: Januari 2001
  • Laatst online: 06-11-2022

JustMe

#1 R@H

Nice flush TCF!!

Cureseekers | There's something wrong with this yogurt


Acties:
  • 0 Henk 'm!

  • My
  • Registratie: Mei 2000
  • Laatst online: 07-06 09:26

My

Morgen neemt TS de eerste plaats over...

(Daily dan hè :P)

Acties:
  • 0 Henk 'm!

  • DPCquatar
  • Registratie: Februari 2001
  • Laatst online: 03-12-2024

DPCquatar

Kwaak

My schreef op 11 november 2003 @ 17:53:
Morgen neemt TS de eerste plaats over...

(Daily dan hè :P)
Dacht het niet.................... >:)

Cruncht tegenwoordig meestal vault projects


Acties:
  • 0 Henk 'm!

  • ColdFusion
  • Registratie: Augustus 2000
  • Laatst online: 30-05 12:47
My schreef op 11 november 2003 @ 17:53:
Morgen neemt TS de eerste plaats over...

(Daily dan hè :P)
Look again 8)

_______________-=Team ColdFusion=-_______________
#2 DPAD - #2 OGR27 - #2 F@H - #3 R@H - #5 RC5 - #6 SoB - #10 WCG
||| #1 ECCp109 - #1 ECC2-109 - #1 OGR24 - #1 D2OL P1|||


Acties:
  • 0 Henk 'm!

  • geintje
  • Registratie: September 2000
  • Laatst online: 30-05 11:43

geintje

=| zeiken met een glimlach |=

My schreef op 11 november 2003 @ 17:53:
Morgen neemt TS de eerste plaats over...

(Daily dan hè :P)
B) de waarschijnlijk van de mogelijkheid dat dat misschien eens gebeurd is morgen geheel NIET aanwezig :+

[ Voor 4% gewijzigd door geintje op 11-11-2003 18:19 ]

________________________Team ColdFusion________________________
#1 DPAD, #1 OGR-27, #2 F@H, #4 RC5, #18 WCG


Acties:
  • 0 Henk 'm!

  • YakuzA
  • Registratie: Maart 2001
  • Niet online

YakuzA

Wat denk je nou zelluf hey :X

Niet zo gemeen doen tegen TS heh :+

Death smiles at us all, all a man can do is smile back.
PSN


Acties:
  • 0 Henk 'm!

  • ColdFusion
  • Registratie: Augustus 2000
  • Laatst online: 30-05 12:47
YakuzA schreef op 11 november 2003 @ 18:15:
Niet zo gemeen doen tegen TS heh :+
Als je dit al gemeen vindt zou ik me maar voorbereiden voor wat we straks met NC gaan doen :Y) 8)7 :P

_______________-=Team ColdFusion=-_______________
#2 DPAD - #2 OGR27 - #2 F@H - #3 R@H - #5 RC5 - #6 SoB - #10 WCG
||| #1 ECCp109 - #1 ECC2-109 - #1 OGR24 - #1 D2OL P1|||


Acties:
  • 0 Henk 'm!

  • geintje
  • Registratie: September 2000
  • Laatst online: 30-05 11:43

geintje

=| zeiken met een glimlach |=

ColdFusion schreef op 11 november 2003 @ 18:18:
[...]

Als je dit al gemeen vindt zou ik me maar voorbereiden voor wat we straks met NC gaan doen :Y) 8)7 :P
B) ach die zien ons toch nu ook wel komen ? of zijn ze nog :z :Z :z :Z

________________________Team ColdFusion________________________
#1 DPAD, #1 OGR-27, #2 F@H, #4 RC5, #18 WCG


Acties:
  • 0 Henk 'm!

  • Z!oN
  • Registratie: April 2002
  • Laatst online: 03-01 11:44

Z!oN

Team Coldfusion

geintje schreef op 11 november 2003 @ 18:13:
[...]

B) de waarschijnlijk van de mogelijkheid dat dat misschien eens gebeurd is morgen geheel NIET aanwezig :+
Met een aan zekerheid grenzende waarschijnlijkheid kan ik dit bevestigen :+

http://www.teamcoldfusion.nl


Acties:
  • 0 Henk 'm!

  • _WgV_
  • Registratie: Februari 2002
  • Laatst online: 24-06 10:47

_WgV_

It's a magical world!

Mja dat betekend iig een zeer chille score voor DPC, maar das wel fijn :D

#StopBurningStuff


Acties:
  • 0 Henk 'm!

  • JustMe
  • Registratie: Januari 2001
  • Laatst online: 06-11-2022

JustMe

#1 R@H

Gaaf dagje :D:D

Cureseekers | There's something wrong with this yogurt


Acties:
  • 0 Henk 'm!

Anoniem: 49384

Lap tekst:
All,
On Saturday, Nov 8th, we had a system failure within our Statistics Processing system. Please let me take a moment to explain.

First, some background and an outline:

Our project is segmented across many servers to provide stability and scale. In version 2.0, processes that all lived together on 1.2 received their own complete server.

When you register a new node, this contacts a Registration Server which will provide a Node ID and create an entry in the appropriate Member database.
When the client begins, it performs a handshake to the webserver to get headlines, graphics, config file, a broadcast message etc. It also makes a request to Scheduler server, asking for Tasks. In days of yore we had problems with this process running out of work to give to the nodes, but now that it's on it's own server - everything here is clean… and a candidate queue size up to 2000 helps smooth over any rough spots.

When the Client completes units of work and performs it's next handshake, it sends the results to the Results Handler. The RH performs several important tasks. First, it validates that the file is intact. If not, it throws it away. Next, it parses the file and creates two files - the first is the analysis file that has only scientific data and is added to the NAS repository known as DRAS (Docking Result Analysis System). This offline system is where our scientists evaluate results and make decisions on direction; the repository contains no member information and is exclusively results. The second file is a small file that contains the node details and the basic details of the results - not the whole file, just enough for stats. The Results Handler then sends the large file to DRAS and then saves the small files to folders sorted by target. At scheduled intervals, a process FTP's these files to the Stats server.

The Stats Server does a lot of work. When it reads a result file, it credits the node with the result. It then checks to see if this is the best result this node has ever done for this Target - updating if so. It then checks to see if this is the best result ever for this Target. Important to note - docking is not a perfect science. We have a reality check that energy must be less than -50.00 or the result will be rejected. The node will get credit for completing the unit of work, but this can not be the "best" anything since it is not realistic. Once we have evaluated the result and credited the node (making notes of any new bests) the Stat processor must update the aggregate files for the node, for the member, and for the team. Once all this is complete, it compiles new files for each grouping and hands it off to another process. The Statistics Viewer takes the complete stat files, loads them into memory, and then updates the webpage. This happens approximately every half hour. When the processing queue is very large, the Stat Processor can not compile a complete file for presentation within the half hour. The web page will continue to show the old file until the Stat Viewer has a new file ready.

OK, that's how it works. Now what happens when it breaks.

If the Scheduler fails, then no one gets new tasks until it comes back online. With a full queue, this could be an entire week before it becomes a problem but everything else will work fine; it's like running out of gas for that node.
If the Results Handler fails, then no one can return results. The Client will receive an error indicating that it could not upload results and they will remain locally until the Result Handler comes online. Again, inconvenient - but nothing gets lost. We could run without a result handler for a very long time but that would defeat the entire purpose of the project. Eventually, stats would freeze since there will be nothing new upon which to report.
If the Stat Processor fails, then the Result Handler can not FTP the result files. These will build up in queue on the Result Handler until the Stat Processor comes online and then the Stat Processor will work through the queue as fast as it can. Nothing is lost, we are just behind schedule.

What happened on Saturday?

On Saturday, we had an technical error where we could not create new files on the Stat Processor. This created two major problems. First, stats were frozen. The Result Handler stopped feeding results since no new files could be created which stopped the problem from getting worse. However there is still problem two - stats in progress. Since a new file could not be created, all the stats in progress would not be written out to the files for the Viewer. These became dropped. We lost the stats from this window of time and can not recover them. All the scientific data is safely stored in DRAS, but it does not contain any of the node info we would need to recreate the stats. Science is safe - but we can not report on the work the community did during that segment of time.

We are working to assure that gaps such as this get closed, providing improved availability and an increased confidence in our whole system, not just Stats. I apologize personally to those nodes and teams impacted by the loss of stats and can only promise to diligently pursue resolution to issues that you raise. Hopefully my outline here improves everyone's overall understanding of the architecture and can create some dialog.

Thanks for everyone's patience and keep crunching!
Helaas dus wat stats kwijt :/ Toch blijf ik erbij dat het project vele malen stabieler is. De fout zoals die nu is voorgekomen zal niet snel nog een keer voorkomen. Charles is zo fanatiek als een jihad strijder :D als het op stabiliteit en projectverbetering aankomt dus ik heb er vol vertrouwen in.

Acties:
  • 0 Henk 'm!

  • Bastien
  • Registratie: Augustus 2001
  • Niet online

Bastien

Probleemeigenaar

Net verhaal... verduidelijkt meteen hoe ze de handel verwerken daar :) Jammer dat het weg is maar goed... het zij zo...

Heeft NC een dagje uitstel >:) }:O

Zolang er mensen zijn zal de wereld nooit duurzaam worden.
Zijn er wel genoeg kliffen op de wereld waar 8 miljard mensen zich als lemmingen vanaf kunnen storten?


Acties:
  • 0 Henk 'm!

  • Mobster
  • Registratie: Februari 2000
  • Laatst online: 07-06-2016

Mobster

Los Alcoholicos

Charles = Afbeeldingslocatie: http://www.eamd.nl/smilies/16.gif

Een samenvatting voor de mensen die lui zijn en/of niet zo goed Engels kunnen lezen? :>
niet voor mij hoor :P

Toch maar eens een andere sig bedenken :P


Acties:
  • 0 Henk 'm!

  • My
  • Registratie: Mei 2000
  • Laatst online: 07-06 09:26

My

Nou ja, TS bijna eerste dan :P

Acties:
  • 0 Henk 'm!

  • geintje
  • Registratie: September 2000
  • Laatst online: 30-05 11:43

geintje

=| zeiken met een glimlach |=

morgen weer een kans.. misschien dan :+

________________________Team ColdFusion________________________
#1 DPAD, #1 OGR-27, #2 F@H, #4 RC5, #18 WCG


Acties:
  • 0 Henk 'm!

  • JustMe
  • Registratie: Januari 2001
  • Laatst online: 06-11-2022

JustMe

#1 R@H

Anoniem: 49384 schreef op 11 november 2003 @ 21:23:
Lap tekst:


[...]


Helaas dus wat stats kwijt :/ Toch blijf ik erbij dat het project vele malen stabieler is. De fout zoals die nu is voorgekomen zal niet snel nog een keer voorkomen. Charles is zo fanatiek als een jihad strijder :D als het op stabiliteit en projectverbetering aankomt dus ik heb er vol vertrouwen in.
Je hebt helemaal gelijk, we moeten dit maar snel vergeten, want ten opzichte van de oude versie gaat het nu echt vele malen beter. Maar wel scherp blijven zodat de kwaliteit gewaarborgt blijft heh lui! ;)

Cureseekers | There's something wrong with this yogurt


Acties:
  • 0 Henk 'm!

  • DPCquatar
  • Registratie: Februari 2001
  • Laatst online: 03-12-2024

DPCquatar

Kwaak

Volgens de laatste berekeningen hebben we over 14 dagen een nieuwe nummer 1
Of komt er nog wat leuks van die rumoerige beesten?

Wordt vervolgd........
Morgen weer een nieuwe aflevering van deze titatenstrijd.


May the best cow win }:O

Cruncht tegenwoordig meestal vault projects

Pagina: 1