Can someone explain

[illusion88]

Title says it all.

 

[ICXRa]

I'm not running BOINC so I don't have al the details, perhaps Rattledagger can chime in here but it is basically SETI 2 with some nice additional features like a built in queue and service install....I think. I believe it will also eliminate team jumping as the WU"s will stay with the team they were produced with but the grand total will always stay with the actual person who crunched them just like it is with F@H. I seem to recall seeing that current BOINC WU's don't count toward anything but this may have changed in the last several months but they have been working the bugs out of it for a while now and I'm not sure they have them all fixed even now.

 

[Rattledagger]

Adding to ICXRa's post, BOINC also have multi-cpu-support and multi-project-support with configurable resource-shares between them.

Variable credit for different wu, meaning "turbo-wu" only using 1 minute to crunch gives very little credit while VLAR gives more credit than normal wu and so on. Some projects also have multiple applications/wu-types, and the crediting is comparable across them and across projects.

BOINC keeps track of who got assigned a "result", and only this user can return this result, and can only return it once.
Credit is only given after result validated, in seti a wu isn't tried validated before got 3 "success"-results.
This means it can take some days before getting the credit, but you're also sure anyone intentionally or unintentionally (like too much overclock) only returns garbage will not get any credit for this garbage.

SETI@home for BOINC have inserted the validated results into the science-database for over a year, so even during half a year of beta-testing the BOINC-version did meaningful science.

The "core client" of BOINC must be manually downloaded and installed, but the science-applications is automatically upgraded. This means, bug-fixes, improvements and new applications is easily upgraded.
Also means can fairly easily start using data from other places than Arecibo.

In seti "classic", every wu have now on average been crunched 8.03 times. :Q and still increasing so probably more like 15.
Under seti for BOINC, every wu is only sent to 4 different users, and tried validated with 3 results, only if errors is the wu sent more times.
This means, the BOINC-version is more than 100% more effective doing science.

 

[illusion88]

A few questions:

• Is it slower then CLi version of SETI? Does it take more time per WU?

• Does it count any differently to TA then SETI classic?

 

[Rattledagger]

Current result-speed is comparable to cli-v3.08, but due to less results per wu than "classic" it really means more science done in BOINC.

They've also got an optimization-project that should improve crunching-speed later on.

BOINC doesn't use "1 wu = 1 credit", but variable credit per wu based on actual cpu-usage.
Therefore, the crediting isn't comparable so everyone starts at zero again under BOINC.
Well, your old "classic"-score will also show up if you can activate your old account, but only as information-message, it doesn't count against your BOINC-credit. The current "classic"-score is from 14. May 2004, this will be updated after "classic" killed off.