SZTAKI Desktop Grid: Project closes on May 31

[RSS]

SZTAKI Desktop Grid is going to close down officially on May 31. We will try to keep the server open as long as possible afterwards if needed, so all in-progress work units can be finished. Next, the project home page will be replaced by a static site containing the description and results for the applications of the project, and will be updated when we get new results from the scientists. We would like to thank you for your support and dedication for SZDG throughout the years!

More... (http://szdg.lpds.sztaki.hu/szdg/all_news.php#151)

[denim]

So long SZDG. You've given me 10 years of low credit return and many errors. Thank you for the 250k milestone. Will slowly move SZDG efforts to Rosetta.

[Bryan]

So long SZDG. You've given me 10 years of low credit return and many errors.

=)) Boy that's the truth!

[denim]

=)) Boy that's the truth!

It's always sad when a project you have been on for so long goes away. SZDG was one of the first ones I started crunching after SETI.

[scole of TSBT]

Still waiting...

[Mumps]

Still have 560 WU's out in the wild, so there's still results to return. :)

[Al]

I haven't gotten any in 24 hours, guess it doesn't like me. But if that's the end of it, if SG got them all they still couldn't pass us on FB.

[Bryan]

I have 7 "fake" accounts getting about 1100 each/day. I'll move them over to SUSA ... been running for Ars Technica to keep HardOCP from passing them for 3rd position :D I will run those accounts until the Sprint starts and then I'm going to shut them down.

EDIT: BTW, I'm still getting work.

[scole of TSBT]

Still have 560 WU's out in the wild, so there's still results to return. :)
There are now 665 WUs in the wild. I'm beginning to think the announcement about shutting down was a ruse to get people to flock to the project.

[Bryan]

There are now 665 WUs in the wild. I'm beginning to think the announcement about shutting down was a ruse to get people to flock to the project.

I think you're right. I'm going to stay on it until right before the Sprint starts and then I'm permanently pulling the plug!

[Al]

For a project that closed the end of May, Sztaki wus continue to go up. At this rate I may make 500k after all, though not likely.

[scole of TSBT]

Results in progress 1,107

[Ron Shurtz]

SG has been slowly gaining on us.
They are now 71K back.
They have been gaining about 1-2 K per day.
They appear to be trying to gain the 7 points for moving into first place before the project really shuts down (If and when it does).
(I wonder if any of the French are helping them.)

[Ron Shurtz]

SG gained another 3K yesterday. :eek:
They are only 68K back now.
I have BOINC Updater running to pester Sztaki every 3 minutes.
There is very little work available but the WUs in progress keep going up.
4855 last time I checked.

[Egilman]

SG gained another 3K yesterday. :eek:
They are only 68K back now.
I have BOINC Updater running to pester Sztaki every 3 minutes.
There is very little work available but the WUs in progress keep going up.
4855 last time I checked.

I would join up and help out and it opened my old account, but won't let me join any team.....

Sorry.

[Bryan]

Ron is correct that SG is gaining on us. It is difficult to say whether the project will continue putting out work, but we certainly don't want to get passed and lose the top position.

SUSA could use your help!

Speaking of help, there are a couple of folks that are within striking distance of milestones:

Sphynx needs 87k to hit the 500k MM
MindCrime needs 27k to reach the 250k MM

You might consider contacting them for a login and give them a boost. I hate to see them win the "I was too busy scratching my butt to get my MM" award. I've collected too many of those myself =))

NOTE: Sztaki does NOT support multiple BOINC instances. However if you have the capability you can attach 1 instance on a machine to an account and then attach a 2nd instance to someone else's account. Set updater or use a tickler script and update each machine at a 2 to 2.5 minute interval.

[Egilman]

Ron is correct that SG is gaining on us. It is difficult to say whether the project will continue putting out work, but we certainly don't want to get passed and lose the top position.

SUSA could use your help!

.....

......"I was too busy scratching my butt to get my MM" award. I've collected too many of those myself =))


I've gotten a few of those myself.....

Update:

I tried again to join the team and the link was there today! So I joined.....

will be helping out if it gives me any work..... Last time I was connected it wouldn't give me any....

Update II:

I've joined on Haf1, I don't have the executables and it doesn't seem to download them.... just sitting on an empty directory, so it really wouldn't know what to do with them if it got some....

[Bryan]

I've gotten a few of those myself.....

Update:

I tried again to join the team and the link was there today! So I joined.....

will be helping out if it gives me any work..... Last time I was connected it wouldn't give me any....

EG, setup a tickler script -or- use Updater and hit each machine at 2 to 2.5 minutes.

[OldChap]

I have been running this on VM's, in my case 28 VM's each allowed 2GB and 20GB of disk, on one 24t rig that has 64GB installed

Running as Grumpy, I am getting quite good numbers when using a tickler: https://stats.free-dc.org/team/sztaki/305

If you can you should over commit the cores running by running something else on the host. This makes the SZ tasks take longer so if you run benchmarks on the vm with nothing else running and then revert to both vm and host running work you should end up with your tasks Claiming more.

I feel that points is always awarded to the lowest claim level so sometimes you can get big awards this way. I think my current average is around 7 pts per with high twenties as my max

Really, is is not a big job to install VM's and this.

If you plan to do a big rig like mine you can get the first one right with all config files done etc then before you start boinc, clone it as many times as you want.

before starting boinc on the others you should rename each VM hostname and attend to the network address.

Once done and SZ is running then maybe add ticklers for Cas and Mindmodelling etc.

Mine is happy to use T 4 on a total of 40 VM's

Can add detail if needed but most of you know a bunch more than I do about this...... but my thought is that if 1 person builds big or everybody runs a few you place is secure

[Bryan]

@OC - haven't done the VM thing in a number of years. You can actually start BOINC and do the benchmarks - just don't attach to any projects. Then you make your clones from that. IIRC there is a setting that needs to be made so that the VM actually changes the Mac address so projects see it as a different machine.

The benefit to that technique versus multiple instances of BOINC is you can do as many machines on a single account as you want. On projects that support multiple BOINC clients you can do the same thing with multi clients.

[Bryan]

@EG, if the project actually decides to give you some work on those machines it will download the executables along with the WU.

[Ron Shurtz]

Results in progress are finally dropping.

Now at only 1283.

Who knows. It may be close to the end or another dip before some more WUs mysteriously appear.

I hadn't checked for a day and a half.
I blew a PSU on my main machine after swapping out the GPU.
I have a new PSU arriving tomorrow.
I'll find out then if that was all that blew.

SG was just making me nervous.
Didn't want to lose 7 FB points just before the project closes for good. :D

[Bryan]

It was a good call Ron. No sense in screwing up at the last minute :D

Hope that PSU fixes everything for you!

[Egilman]

EG, setup a tickler script -or- use Updater and hit each machine at 2 to 2.5 minutes.

The updater is hitting it at 125 seconds per, nothing yet.... (and it's been three hours)

So I'll patiently wait it out....

In the meantime it'll be running rakesearch at 0% just to keep it busy