SETI@Home (BOINC)

[Mayabird]

Yes, fellow fans of distributed computing, it's finally up and running! You can download it any time between now and whenever the SETI people tell you the original one is done.

Mine's already hard at work.

http://setiboinc.ssl.berkeley.edu/sah/

[Vohu Manah]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

[InnocentBystander]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

The search for life beyond our stinky little clump of rock is quite important, acutally.

[Lagmonster]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

The search for life beyond our stinky little clump of rock is quite important, acutally.

Not nearly as important is improving life on our stinky little climb of rock, mind you.

[Pu-239]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

http://folding.stanford.edu

I'm running it on my mostly idle webserver.

[darthdavid]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

http://folding.stanford.edu

I'm running it on my mostly idle webserver.

What's the sd.net Folding Team? Do we have one?

[Pu-239]

I may have to reconsider participating in a distributed computing project. Though I'd prefer something other than Seti@Home. Know of any distributed computing project related to like.. curing cancer or something?

http://folding.stanford.edu

I'm running it on my mostly idle webserver.

What's the sd.net Folding Team? Do we have one?

I don't think so.

[Mitth`raw`nuruodo]

Could someone explain to me what the difference between this and the version of SETI I used 5 years ago is? I'm ignorant.

[Drooling Iguana]

Scientific progress goes BOINC?

[Vohu Manah]

http://folding.stanford.edu

I'm running it on my mostly idle webserver.

They're still around!?! Already restarting with them.

[Symmetry]

What's the sd.net Folding Team? Do we have one?

I'm already on another team.

[Mayabird]

Could someone explain to me what the difference between this and the version of SETI I used 5 years ago is? I'm ignorant.

For one thing, the old SETI is set to end in the not-too-distance future, while this new one will be lasting for quite a while longer.

BOINC isn't just for SETI (although I don't know of any other projects yet that is using it). If the protein folding project or some others started using it, you could allocate resources for all your various projects. Protein folding 50%, climate modeling 25%, and SETI 25% for example. Also, you can determine different things such as how large a file you can get (you don't want to download 10 MB files to scan if you're on dialup, after all). There's also a different method of determining the credit you get from doing the work.

There are plans to start processing data from the southern hemisphere (currently all of it is from the Arecibo radio telescope in Puerto Rico) as well as going over all the previous SETI info but using different methods. I'm not sure if those are up yet, so what I'm doing is probably just a continuation of the previous project, just with me downloading eight at a time without any special add-ons.

Also, the screensaver is now 3-D. Everyone loves 3-D graphics.

And that's all I know. I'm lazy and easily mesmerized by pretty colors, so I haven't done much more research.

[Pu-239]

What's the sd.net Folding Team? Do we have one?

I'm already on another team.

You can belong to multiple teams, and I believe your points are duplicated for all teams.

[Uraniun235]

*remembers his oft-forgotten Folding@Home client, fires it up*

[Rob Wilson]

Yes, fellow fans of distributed computing, it's finally up and running! You can download it any time between now and whenever the SETI people tell you the original one is done.

Mine's already hard at work.

http://setiboinc.ssl.berkeley.edu/sah/

Mines been going for 3 days now... and in that time has managed to tell me that there's no work to do for 90% of that time. However it did do three units of work. They still don't show in my account and i have no credit for them. What the fuck is going on with that?

[Datana]

Mines been going for 3 days now... and in that time has managed to tell me that there's no work to do for 90% of that time. However it did do three units of work. They still don't show in my account and i have no credit for them. What the fuck is going on with that?

It's part of the new verification scheme. To keep users from cheating and returning a huge pile of fake results (as happened with the original SETI@Home), you only get credit if someone else has done the same workunit and achieved the same result as you. At that point, the lower of the two claimed credits is assigned to both people (say, your client claimed 20.0 credits of processor work and the other's 35.3, you both get 20.0). Thus, there's often a lag of several days between completing a workunit and getting credit if you have a fast system (as the other guy with the same workunit is still crunching away at his/her copy).

As for no new work, it seems to be a recurring problem with BOINC right now. Hopefully it'll sort itself out soon.

[Rob Wilson]

Mines been going for 3 days now... and in that time has managed to tell me that there's no work to do for 90% of that time. However it did do three units of work. They still don't show in my account and i have no credit for them. What the fuck is going on with that?

It's part of the new verification scheme. To keep users from cheating and returning a huge pile of fake results (as happened with the original SETI@Home), you only get credit if someone else has done the same workunit and achieved the same result as you. At that point, the lower of the two claimed credits is assigned to both people (say, your client claimed 20.0 credits of processor work and the other's 35.3, you both get 20.0). Thus, there's often a lag of several days between completing a workunit and getting credit if you have a fast system (as the other guy with the same workunit is still crunching away at his/her copy).

As for no new work, it seems to be a recurring problem with BOINC right now. Hopefully it'll sort itself out soon.

People were returning fake units?

And what happens in this system if the units I'm doing are being done by 1 of the near 90% of people that don't bether to stay the course. IE, don't even complete their units? I'm then screwed for credits, but most importantly any results i return are completely ignored (so even if I get something interesting in my results, they'll be unused by the project because no one has confirmed them.)

[Datana]

People were returning fake units?

And what happens in this system if the units I'm doing are being done by 1 of the near 90% of people that don't bether to stay the course. IE, don't even complete their units? I'm then screwed for credits, but most importantly any results i return are completely ignored (so even if I get something interesting in my results, they'll be unused by the project because no one has confirmed them.)

As for fake units, yeah. Especially in version 2.xx, there was rampant cheating for more workunits completed (there were several exploits that could be easily performed to get a result returned before processing was complete, and there were even means of autogenerating fake returns). The most recent versions of the regular client has additional safeguards in place, so it's not as critical an issue as it was before, but it seems that the SETI folks want more checks and added them in BOINC.

On the issue of not being assigned credit due to an abandoned unit, the new workunits expire if you haven't completed them by a particular date (usually a week from the day you recieve the file). If not returned by the date assigned to it, the BOINC server reassigns the unit to someone else.

[Rob Wilson]

As for fake units, yeah. Especially in version 2.xx, there was rampant cheating for more workunits completed (there were several exploits that could be easily performed to get a result returned before processing was complete, and there were even means of autogenerating fake returns). The most recent versions of the regular client has additional safeguards in place, so it's not as critical an issue as it was before, but it seems that the SETI folks want more checks and added them in BOINC.

On the issue of not being assigned credit due to an abandoned unit, the new workunits expire if you haven't completed them by a particular date (usually a week from the day you recieve the file). If not returned by the date assigned to it, the BOINC server reassigns the unit to someone else.

Ah, ok cheers for clearing that up.

Now I just hav to workout why the BOINC account page shows the wrong number of SETI@home classic units completed?
The BOINC page says 112, the ASVS group page shows 121, and has done since May, before BOINC appeared.

[Defiant]

As for fake units, yeah. Especially in version 2.xx, there was rampant cheating for more workunits completed (there were several exploits that could be easily performed to get a result returned before processing was complete, and there were even means of autogenerating fake returns). The most recent versions of the regular client has additional safeguards in place, so it's not as critical an issue as it was before, but it seems that the SETI folks want more checks and added them in BOINC.

On the issue of not being assigned credit due to an abandoned unit, the new workunits expire if you haven't completed them by a particular date (usually a week from the day you recieve the file). If not returned by the date assigned to it, the BOINC server reassigns the unit to someone else.

Ah, ok cheers for clearing that up.

Now I just hav to workout why the BOINC account page shows the wrong number of SETI@home classic units completed?
The BOINC page says 112, the ASVS group page shows 121, and has done since May, before BOINC appeared.

When they began the initial setup of BOINC, they took a snapshot of your units as of May 15th. They said that they'll do another snapshot just before they shutdown the Classic SETI@home.

[Rob Wilson]

Now I just hav to workout why the BOINC account page shows the wrong number of SETI@home classic units completed?
The BOINC page says 112, the ASVS group page shows 121, and has done since May, before BOINC appeared.

When they began the initial setup of BOINC, they took a snapshot of your units as of May 15th. They said that they'll do another snapshot just before they shutdown the Classic SETI@home.

Ah, ok then. Cheers for the info.