Technical News and SETI@home daily stats for 28.09.2007.

[Rattledagger]

I think we've figured out what the real issue with gowron (the NAS box that holds the work units) is. It's that those who don't learn from history are doomed to repeat it. In other words we had this problem once before.... about 7 years ago. Yesterday, David reminded me how much smarter I was back then.

We have gowron configured into RAID 5 arrays, which is positively stupid for the job it's doing. RAID 5 works great for reading large files since you are reading from (in our case) from 5 drives at once. For reading small files like work units it sucks, because every time you read one you are moving the heads of all 6 drives and then reading a small amount of data. In other words we're spending all of our time seeking and very little actually reading.

A much better way would be to have 6 drives set up as a concatenation of three RAID 1 mirrors. That way each pair of drives would seek individually, so we'd reduce our relative time spent seeking by a factor of 3 at the expense of slowing down our data reads by a factor of 2.5. It's worth the trade, so I'm moving workunits off of gowron so we can reconfigure.

I don't think solid state (flash) is a good fit for us because we rewrite all of our data fairly constantly. I think we'd hit the lifetime rewrite limit of NAND flash in a couple years.

Eric

#____Total Work Done____Todays WD_______AWD________overtake________Team-name
01______338.522.314______511.130______566.037______impossible______SETI.USA
02______276.802.333______358.966______362.590______impossible______SETI.Germany
03______119.139.153______216.866______214.339______impossible______L'Alliance Francophone
04______110.541.216_______50.178_______63.072______impossible______BroadbandReports.com Team Starfire
05______101.208.062______152.852______157.735______impossible______BOINC Synergy
06_______92.355.098_______85.618_______80.895______impossible______Czech National Team
07_______82.797.777______107.326______107.258______impossible______SETI@Netherlands
08_______68.569.335______154.946______152.528______impossible______The Knights Who Say Ni!
09_______43.959.574______-21.346______-15.615______2.815 days______OcUK - Overclockers UK
10_______37.761.793_______50.951_______50.969______impossible______Team Art Bell
11_______36.567.570_______17.314_______21.313______impossible______Overclockers.com
12_______33.099.200_______25.684_______36.732______impossible______Team 2ch
13_______31.190.539______-42.906______-35.403________881 days______BOINC.Italy
14_______28.290.752_______49.175_______44.741______impossible______The Planetary Society
15_______25.234.449_______76.718_______69.977______impossible______Team MacNN
16_______17.339.602______-12.998_______-4.841______3.582 days______Ars Technica
17________4.572.818______-22.080______-11.398________401 days______Universe Examiners
18__________543.707_______33.014_______32.183______impossible______SETI@Taiwan
19__________370.848________7.698_______16.855______impossible______Phoenix Rising
20_______59.395.492_______96.408_______92.269______notanoption_____TeAm AnandTech
21_______-4.333.603_______70.563_______65.574_________66 days______Team China
22_______-5.157.096_______-5.144________5.050______1.021 days______Team Starfire World BOINC
23_______-9.726.836______-14.577______-16.015______impossible______Hewlett-Packard
24______-10.926.172________1.045________2.911______3.753 days______Canada
25______-10.949.732______-14.323______-11.625______impossible______Amateur Radio Operators
26______-10.977.787______-11.847_______-5.076______impossible______Dutch Power Cows
27______-11.480.234______-39.988______-32.486______impossible______PC Perspective Killer Frogs
28______-16.347.758______-27.169______-24.192______impossible______Team MacAddict
29______-16.690.794______-66.069______-59.197______impossible______Planet 3DNow!
30______-16.851.507______-49.864______-43.959______impossible______2CPU.com
31______-17.242.514______-33.509______-28.971______impossible______Team NIPPON
32______-18.907.928_______-2.253________5.454______3.467 days______UK BOINC Team
33______-19.270.609______-14.380______-12.936______impossible______BOINC SETI@home RUSSIA
34______-19.543.317______-24.859______-17.956______impossible______BOINC@Denmark
35______-19.652.217________3.470_______14.166______1.387 days______BOINC@AUSTRALIA
36______-20.021.474______-23.689______-22.093______impossible______Hungary
37______-20.748.483_______-8.780_______-9.661______impossible______US NAVY
38______-22.464.467______-50.016______-50.163______impossible______Portugal@Home
39______-23.582.297______-10.491_______-5.603______impossible______U.S.Air Force
40______-25.653.737______-40.029______-34.972______impossible______SETI.hr
41______-25.977.210_______57.282_______57.157________454 days______AUSTRIA - NATIONAL - TEAM
42______-26.505.570______-80.702______-75.930______impossible______Picard
43______-26.733.401______-43.697______-38.350______impossible______SETI@klamm.de
44______-27.287.163______-76.300______-71.081______impossible______LittleWhiteDog
45______-28.076.368______-54.964______-49.348______impossible______Team EDGE
46______-28.436.231______-52.891______-47.841______impossible______HispaSeti & BOINC
47______-30.331.364______-48.723______-45.121______impossible______SETI Sverige [Sweden]
48______-30.393.980______-24.871______-20.536______impossible______BOINC.SK
49______-32.432.725______-62.723______-61.088______impossible______BOINC UK
50______-32.990.765______-58.719______-51.690______impossible______PurePC.pl

Appart for Anandtech's stats, it shows how much more/less than Anandtech.
Also shows based on Average Work Done how many days for Anandtech to overtake the team, or be overtaken by a team behind

 

[mk]

Thank you for the update, Rattledagger. :beer:

 

[Zim Hosein]

Thanks for the SETI@home daily stats and tech news Rattledagger! :beer:

 

[mk]

Originally posted by: Zim Hosein
Thanks for the SETI@home daily stats and tech news Rattledagger! :beer:

Hi there Zim.

People have been missing you.

 

[Assimilator1]

Thanks RD ,sounds like they'll be removing a major bottleneck