How hard can this be?


log in

Advanced search

Message boards : SETI@home Science : How hard can this be?

Author Message
robbie
Send message
Joined: 18 May 99
Posts: 1
Credit: 40,291
RAC: 0
United States
Message 21995 - Posted: 3 Sep 2004, 20:29:19 UTC

How come Predictor and Climateprediction can get this working, but Seti is having numerous, NUMEROUS problems? In my line of work (IT) this long of an outage would mean some heads would roll. I am only back at Seti because Predictor is upgrading.

Profile Mr. KevvyProject donor
Volunteer tester
Avatar
Send message
Joined: 15 May 99
Posts: 714
Credit: 74,567,999
RAC: 73,806
Canada
Message 22011 - Posted: 3 Sep 2004, 20:53:14 UTC - in response to Message 21995.
Last modified: 4 Sep 2004, 2:14:54 UTC

> How come Predictor and Climateprediction can get this working, but Seti is
> having numerous, NUMEROUS problems?

ClimatePrediction.net = 6,537 PC's
Predictor@Home = 13,609 PC's
SETI@Home = 83,381 PC's

( from http://www.setisynergy.com/stats/index.php )

I would imagine it's a matter of scale, and that some of the outages were due to the upgrades to prevent even more problems when the *450,000* SAH Classic participants switch over. Ouch.

Angstrom
Send message
Joined: 20 Sep 99
Posts: 205
Credit: 10,131
RAC: 0
United Kingdom
Message 22020 - Posted: 3 Sep 2004, 21:10:48 UTC

I think it is a largely a matter of scale as Mr Kevvy says. Also probably a lack of hardware resources. The other BOINC projects are also benefiting from the fixing of code, db's etc that is going on here as those fixes can then be immediately implemented elsewhere.

Like you I am back from Predictor. For me the problem here isnt that there are long outages but rather that the flow of informtion from the Seti team is next to non-existant.


Neil

Heffed
Volunteer tester
Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 22167 - Posted: 4 Sep 2004, 3:45:19 UTC

The S@H team is not only trying to keep the hardware going, but develop the software as well...

I guess my answer to your question would be, hard.

Profile Everette Dobbins
Send message
Joined: 13 Jan 00
Posts: 263
Credit: 6,102,098
RAC: 0
United States
Message 22190 - Posted: 4 Sep 2004, 6:05:05 UTC - in response to Message 22011.

> > How come Predictor and Climateprediction can get this working, but Seti
> is
> > having numerous, NUMEROUS problems?
>
> ClimatePrediction.net = 6,537 PC's
> Predictor@Home = 13,609 PC's
> SETI@Home = 83,381 PC's
>
> ( from http://www.setisynergy.com/stats/index.php )
>
> I would imagine it's a matter of scale, and that some of the outages were due
> to the upgrades to prevent even more problems when the *450,000* SAH Classic
> participants switch over. Ouch.
>
I hope they dont switch over seti classic for at least a another year. 4-5 times a week or more when BOINC craps out I run seti classic.

Alex
Send message
Joined: 26 Sep 01
Posts: 260
Credit: 2,327
RAC: 0
Canada
Message 22211 - Posted: 4 Sep 2004, 8:10:00 UTC - in response to Message 21995.




Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8452
Credit: 4,151,898
RAC: 1,761
United Kingdom
Message 22268 - Posted: 4 Sep 2004, 14:36:02 UTC - in response to Message 22211.

> [ v4.03 CPDN run Fortran crash on a Windows machine ]

Yes... Very good... Now upgrade to v4.06.

This CPDN problem is very quickly being worked on. Development continues.


And your point was?

Get crunchin',
Martin

Profile bjacke
Volunteer tester
Avatar
Send message
Joined: 14 Apr 02
Posts: 346
Credit: 13,761
RAC: 0
Germany
Message 22269 - Posted: 4 Sep 2004, 14:53:34 UTC

Thx for the Info. CPDN I'm comeing once again ;-).

The whole is more then the sum of its particles.
Aristoteles
Best wishes from Berlin(52°35'N,13°23'O), Basti

Alex
Send message
Joined: 26 Sep 01
Posts: 260
Credit: 2,327
RAC: 0
Canada
Message 22310 - Posted: 4 Sep 2004, 17:36:48 UTC - in response to Message 22268.
Last modified: 4 Sep 2004, 17:40:24 UTC

> > [ v4.03 CPDN run Fortran crash on a Windows machine ]
>
> Yes... Very good... Now upgrade to v4.06.
>
> This CPDN problem is very quickly being worked on. Development continues.
>

Running 4.06 wasn't the magic bullet.
The problem was a bad file on a mirrored server.

Ironically, the 'Jump to conclusions' method of problem solving (like your 'just run 4.06' comment) is the kind of problem solving technique that results in the server being down for long periods of time with wrong 'fixes' applied.

Seti/Boinc has been under load since June ('ish). They've had some problems with their system. They've diagnosed multiple things as 'the problem'.
Board users have made their own wild speculations as to what the problems at seti may have been.

>
> And your point was?
>

The point is that the thread starter praised climateprediction and predictor (climate prediction having Fortran errors, and Predictor not even upgrading to 4.x) as being such great projects and criticized Seti.

The whole 'Try something, and if it works, then you've found the magical fix' technique is flawed, because the thing you changed may not have been the item causing the error, and the fact that the system 'works' after the change was just a coincidence.

The 'just run 4.06' response is about as useful as the message board posters who felt that climate prediction's fortran errors were due to :
- overclockers.
- win98
- crappy HP drivers interfering with everything.
- 4.05 Boinc.
- Not running in Win2000 compatiblity mode.
- seti at home user backgrounds.
- Firewalls
- service pack 2.

Message boards : SETI@home Science : How hard can this be?

Copyright © 2014 University of California