DF ending.... question(s) about SETI

[RC Agent]

I'm looking at the MURC team page for SETI.
What's the difference (if any) between:

http://setiweb.ssl.berkeley.edu/team_display.php?teamid=30253

and

http://setiathome2.ssl.berkeley.edu/stats/team/team_25678.html

[Rattledagger]

The 2nd is for "classic", this will be replaced by the BOINC-version later on, so classic will hopefully end this year...

http://setiweb.ssl.berkeley.edu/ is the new BOINC-version that is taking over, and was released 22.06.2004 after long beta-test.
BOINC is a multi-project-platform, there you can join multiple projects and control how much of the time each should crunch. There's also built in caching & multi-cpu-support.

Since seti still can have smaller and sometimes longer server-outages, it's highly recommended to join atleast one other project also.
Appart for seti@home, the other released projects are:
http://climateprediction.net predict weather, with wu taking around 1 month on fast machines, there's no risk of running out of work.
http://lhcathome.cern.ch/ help build the new particle-accelerator at CERN. There's different-sized wu, there the longest is taking around 12 hours.

Other projects, like Astropulse that also will be run by Berkeley, is currently being developed.
http://predictor.scripps.edu/ is currently in partially open alpha, but till Dell manages to deliver the new servers this project is offline...

Also, seti for BOINC is much more efficient than "classic", since a wu is only delivered to 3 different crunchers and only if needed to more, while in classic there's 7,23 results/wu...

[High_Jumbllama]

If a machine runs out of work for one project due to outages, will it then direct the unused resources to the other project?

[Rattledagger]

Yes, if one project isn't handing out work the slack is taken up by another project.

But there's a slight problem in current client, so if you've got multi-cpu & shares with CPDN, make sure to get one wu per cpu. If the CPDN-share is less than 50% you'll normally only download 1 wu. To fix this, either "attach" to CPDN 1st, or for a short while increase the CPDN-share till downloaded one wu per cpu.

[Jon P. Inghram]

Too bad the latency and bandwidth of the net is too poor, I could probably use some extra CPU and GPU cycles when HL 2 comes out.

[RC Agent]

Thanks for the reply. I've signed up for SETI@home BOINC version and joined team Matrox Users.