Preparing for the XtremLab leg of the BOINC race

[mitro]

I just attached XtremLab here on my work computer and I'm only able to get a handful of WUs at a time. I've got my preferences set for "connect to network every 5 days" and it makes no difference. This is a problem for my work computer because it doesn't have an internet connection all the time (on dial-up) so I need to get WUs pretty much on my schedule.

Am I missing something or is that just how XtremLab works?

[Strongbow]

I was just about to post the exact same question! :?

EDIT UPDATE : From the XtremLab site they state:

"Our application contacts briefly and at regular time our server to obtain the exact time: this implies that our client performs network activity. (even if disabled in BOINC client)..."

[NeoGen]

If you press update again, will it fetch more? If it does, then you only have to do it several times until you got the queue full.

There was also a trick on many projects that was suspending all the workunits one by one and then pressing update. Boinc would download another batch of them. You repeated the process until boinc would complain that it won't get more because it won't finish the work in time.
Then you un-suspend the workunits and let them run. Takes a bit more but also does the job.

[Strongbow]

We can't do that either NeoGen, I've tried that, it defers communication for the amount of time it requires to do all the sent wus. You basically get a message saying "05/05/2006 17:45:30|XtremLab|Message from server: Not sending work - last RPC too recent: 736 sec"

[mitro]

XtremeLab will only allow you to update every 40 minutes or something like that and after 40 minutes I get about 10 WU and those are crunched within 30 minutes leaving the computer idle. :(

EDIT: I may have exaggerated. I have enough now to keep the project working until the next update, but nowhere near what I need over a weekend (or overnight even).

[NeoGen]

Damn... 10 workunits take me over an hour to crunch, so that's why I never saw it happening. That and me keeping a small queue.

It takes a contest for us to find out all these little quirks on the projects... :?

Suggestions anyone? This is getting too close to the hour... should we bump up SZTAKI then?

[mitro]

I'm OK with staying with XtremeLab. You won't run out of work if you are always connected. I"ll just leave my work computer on Leiden, but for others it may matter.

[Electabots]

Oh yea guys, just incase some of you didn't know all the xtremelab wu's take 10 minutes regardless of how fast/slow your cpu is. Might be helpful when trying to plan out how many wu's you need to queue up.

[Lagu]

Hi Guys. Here is another Guy as have the same problems.

I can’t get any WU´s from XtremLab. I have use Boinc 1 time when submitting tasks to Leiden earlier today from Leiden eiter and after that I never got any new work. I have set the contact to the server to 0.1 days and lather change to 5.0 days and after that I restarted Boinc. The Boinc manager sent this message:

2006-05-05 20:15:17|XtremLab|Sending scheduler request to http://xw01.lri.fr:4320/xtremlab_cgi/cgi
2006-05-05 20:15:17|XtremLab|Reason: Requested by user
2006-05-05 20:15:17|XtremLab|Requesting 432000 seconds of new work
2006-05-05 20:15:22|XtremLab|Scheduler request to http://xw01.lri.fr:4320/xtremlab_cgi/cgi succeeded
2006-05-05 20:15:22|XtremLab|Message from server: Not sending work - last RPC too recent: 2375 sec

If I set it to 0.1 day I only got a small amount of work. I want to queue up more work.

Have you guys any advice?

Lagu :evil:

[Lagu]

NeoGen

No more changes. We can´t play here and there. We run as we have said.

Lagu :)

[Empty_5oul]

i think we should stick with it. Its not a major problem.

[Lagu]

Hi

I should download XtremLab´s WU to mu AMD Athlon 1001. I got 3!! work units (Shit) then the server said "Thank you" and close the communication. Should I :cry: or :lol: ?

Now I run QMC on my AThlon 64 who is nearly done and I even have 7 WU´s from uFluid of 5.52 hours each from before the race so the computer not need to go idle.

On my AMD Athlon 1001 I have 3 wu`s from uFluid who take 18.45 Hours each. I got the largest WU´s to my slowest Athlon. These WU´s are before the race too.

Lagu :roll:

[Lagu]

Hi

From XtremLabs forum. This guu have first problems and then he found a solution but he never told the other what he had done. That is to be an egoistic one.

Well, today I downloaded my firts 10 WUs. All are computed without any error.
But now I will download new work, but everytime I got the message that my last RPC was too recent.
Then the communication is delayed for 40 minutes. And within these 40 minutes I will get no new work. Why?

What the hell means [last RPC was too recent.] I have got the same but I hav enever seen it before. Is Boinc sick or stjupid? It is me Lagu as is asking.

I only will crunch!:(

Have found a solution :) Not me!!

[gamer007]

So what is going on? I'm glancing around, I see errors. Is it still gonna be XtremLab?

If I don't get an answer soon, I'm assuming and will start caching.

OMG your kidding. Xtremlab thinks every announcement to the server from your computer is a RPC thing..... Even other projects don't think of everything need to be 40mins apart! :evil:

[NeoGen]

I'm open to suggestions... I'm learning quite a few things I didn't know about XtremLab today, which is not good.

If all workunits take 10 mins regarding of the machine power, it's a bit bad for a contest... :?

But since it's a bit too late to change already, I guess the best we can do is keep on going and later ponder about invalidating this race?

[AMDave]

So their project is working exactly as stated. Bit of an improvement over the alpha version.

I spat chips when the previous incarnation of the project lost 3 months worth of my efforts (and everyone else's). But now that I look more closely, they wiped that old slate clean, obviously put their heads together, and now have seriously improved the credit awarding formulae.

I still wish I could get those credit back though. ho hum.
Getting over it and moving right along ... :lol:

The target for the Xtremelab project is to even the playing field for all types of cpu. Their nirvana would be if every WU returned was awarded 1.0 credit. Then they will have accounted for all (almost all) of the differences and can apply those to BOINC to get the crediting systems fixed up once and for all.

Looking at the credit awards in detail, the project are getting waaay closer to this than they used to. Good job in progress this time, I think. (I hope they have a backup system in place now)

If this is going back into the develoment of BOINC then I am all for it.

[AMDave]

BTW
this is a good project for those of us with older / slower CPUs
it is the number involved that really count
You have to be in it to win it ;)