[Merged] Team Seti & distributed computing

[Marge]

I'm still crunching

Originally Posted by Debsy42

I'm still crunching

I didn't know you even took part!

[Richard M]

I've moved:
http://www.cableforum.co.uk/board/sh...postcount=1536

[Maggy J]

Originally Posted by MovedGoalPosts

Are you on the original (classic) version or Boinc?

I think many of the CF team have given up, or jumped ship. There may be a stalwart or two still hanging in there.

Looking at the stats on http://www.setisynergy.com/stats/tea...96&project=sah it seems that swoop101 is the only member returning credits in the last week. (that's for the boinc project)

Boinc.
__________________

Originally Posted by Richard M

I've moved:
http://www.cableforum.co.uk/board/sh...postcount=1536

I don't run my PC all the time so CPU usage of electricity is not a problem.

Well perhaps we/you should amend the Cable Forum team? So it truly represents who is still doing it.Mind with all the helpful info in this thread it would be a shame if it was lost.

Well SETI Classic is being closed soon, and I wonder what will happen to all the people taking part in it.

Some of them have already said there not going onto BOINC, but joining CPDN and others have said they are quiting altogether.

Looks like they don't want to join the BOINC version

[Rob]

There are only a finite number of people who will ever run a distributed computing project. Seti classic being, as far as I'm aware, the firts, will have stirred the imagination and so gained the support of many. As time moved on, other projects have come online, and many of the seti classic users will have moved away.

The advantage of boinc, but also it's disadvantage, is that you can support more than one project at a time. Tha means projects that woudln't have been frist choice get some support, but the first choice projects loose some power.

Classic was popular since it was percieved as secure. You downloaded the classic software and so only the individual WU data then cam and went. No changes were made to software without your consent. Boinc apart from the core client (currently 4.13 - but 4.19 is needed if you use Predictor or Einstein) downloads the software needed for each project automatically and thus that can be changed by each project whenever they like. For many who like to be certain of what is on their PCs that is seen as an issue.

Other projects have turned off their original systems as they have upgraded, others are winding their originals down. It's a natural process of evolution. I suspect the real loss to seti would be those who put classic on all those months or years ago, and have just forgotten it's there crunching merrily away - especially machines that were borged.

[greencreeper]

Originally Posted by MovedGoalPosts

Classic was popular since it was percieved as secure. You downloaded the classic software and so only the individual WU data then cam and went. No changes were made to software without your consent. Boinc apart from the core client (currently 4.13 - but 4.19 is needed if you use Predictor or Einstein) downloads the software needed for each project automatically and thus that can be changed by each project whenever they like. For many who like to be certain of what is on their PCs that is seen as an issue.

Yes - but their paranoia doesn't stand up to scrutiny. We all use software and have absolutely no idea what the software does. How many know that Word documents contain personal information about the author and revision history? To me, the anti-BOINC lobby are just the usual people who resist change because it's scary.

[Rob]

Originally Posted by greencreeper

To me, the anti-BOINC lobby are just the usual people who resist change because it's scary.

But computing is scary, especially if you think they are out to get you

[Maggy J]

I just want to know how long one has to wait for one credits to be acknowledged.Mine seem to be disappearing down a black hole.

Is it day's?Weeks?Months?Years? and when are they going to turn on the review pending credits again?

[greencreeper]

It depends on how many similar results need to be returned for the work unit to be marked as completed, and how fast those with the work unit return their results. So if say a project needs 3 similar results and Bert on his trusty P2 has one of them, you could be in for a long wait

If you look at the server status (Seen here: http://setiweb.ssl.berkeley.edu/sah_status.html) it shows you how many WU's are still being validated for credit. Now everytime that page is update I see the number going down, so it looks as if the validater is catching up (probably due to the new, more powerful server they have running the database).

Always good to look at that page if your getting connection troubles, as they will turn red and show "Not running" if they have crashed, or are shut down for work. Splitters 2 and 3 are usually off, since number 1 can maintain enough WU's. Strange why for the past week all 3 have been running

[Rob]

Seti boinc over the last few weeks had some serious problems with their hardware. As a result altough WUs were being issued (almost normally), crunched and returned, there was then an ever growing queue of unvalidatated WUs. As a result very few people were getting much credit awarded. Your work effort wasn't lost, just severely delayed.

About a week ago seti finally managed to migrate the project to new hardware (what they were running on was almost museum pieces, astounding for such a large project, albeit one without a lot of funding). They had tried to migrate before, but kept having problems.

Now the Validation process is starting to catch up with the backlog. Borkeley's Server Status Page shows the current delay. This did show well over 250,000 WUs at the beginning of the week

This chart which graphically tracks the relative progress of the top two seti teams, illuistrates the difference in amount of WUs before and after the server upgrade. It shows the extent of the backlog, given that project clients have been crunching away almost without interruption.

I doubt borkeley will turn on the pending credit account display. They just have too many for their server to deal with, which adds to another overload.

Just keep crunching. You are doing it for the science, aren't you. The credit is just the icing on the cake.

I have 3 WU's on my laptop that won't update due to a log problem. Scheduler is responding, but not to WU uploads.

The validater has about 16,000 WU's to validate at the moment, and it also looks as if they have shut off the splitters due to the data server crashing (AGAIN!) this morning. Plenty of WU's ready to send though, about 500,000

Had best keep retrying the uploads

[Stuart C]

Originally Posted by greencreeper

Yes - but their paranoia doesn't stand up to scrutiny. We all use software and have absolutely no idea what the software does. How many know that Word documents contain personal information about the author and revision history? To me, the anti-BOINC lobby are just the usual people who resist change because it's scary.

My problem with Boinc is that it is not properly sandboxed (the project software it downloads has the same access to the system that the user that runs it has). BTW, that bit of info comes from the BOINC homepage.

From http://boinc.berkeley.edu/security.php

Intentional abuse of participant hosts by projects

BOINC does nothing to prevent this (e.g. there is no 'sandboxing' of applications). Participants must understand that when they join a BOINC project, they are entrusting the security of their systems to that project.

The other problem is that if you are concerned about individual software packages accessing the network, you may be able to install a firewall and tweak it so that program can't access the net (assuming it does not need to). AFAIK, you cannot do that with BOINC, as any network access appears to come from the BOINC executable, and if you block that from the network, there is no point in running it.

[Maggy J]

Scott?What was wrong with being Scott?

Can't keep up with the name changes around here...mutter,mutter,mutter.....