Asteroids@home - new to me! Whose crunching it?

[networkman]

It's an Intel Xeon 2.8 [Family 15 Model 4 Stepping 8]

It's my main work station[Dell Precision 470], so it has to share time with productivity related items at least 8 hours a day.

 

[Neurodog]

Hey Team!

I'm back from my summer hybernation! Hope everyone is doing well.

I've started crunching again with my home fleet.

Watch your backs.

 

[biodoc]

Bring it on Neurodog!

Welcome back! :thumbsup:

 

[networkman]

@ Assimilator1

Okay, I managed to get the router back up and running, so the test lab is again operational.

Have added a pair of Dell PowerEdge 1850 servers - each is a Dual Xeon 2.8 (displays as 4 CPUs per box) - [Family 15 Model 4 Stepping 3]. One of the servers I expect will be full-time, the other will have duties as a "Minecraft" server for an upcoming Youth program we're testing, so while it'll have some initial output, afterward, it may be sporadic.

 

[Assimilator1]

It's an Intel Xeon 2.8 [Family 15 Model 4 Stepping 8]

It's my main work station[Dell Precision 470], so it has to share time with productivity related items at least 8 hours a day.

Work gets in the way of so much!

Re Xeon, yea like I said those numbers mean nothing to me, I'll Google 'em.
Ah it's a Pentium 4 Prescott ('Pentium 5' ) core, now I know what it is

Sounds like you'll be upping your output a lot then!

Neurodog
Welcome back, and as Biodoc said, bring it on!
What ya crunching with?

 

[networkman]

Looks like our favorite Asteroids projects is having problems again.

None of my rigs have been able to connect to upload, and I was going to add another Xeon E5540 @ 2.53 GHz to the project but it cannot connect to authorize or download tasks. No notices sent to any of the clients.

 

[Sunny129]

fortunately the one host i currently have on the project has ~2 days of work cached. i'm sure the project will be back up within the next few days, maybe sooner. if this outage were due to the server upgrade, we would have gotten a message from the project about it first...

 

[networkman]

I agree. Which is why I'm speculating that it's another unplanned outage. I had set each of my rigs to keep 2 days of work cached, so hopefully that will be sufficient.

 

[GLeeM]

The planned server outage was for Friday

 

[zzuupp]

The planned server outage was for Friday

That's what I thought as well. Mr. Server gets a long weekend.

 

[Sunny129]

according to Kyong, the unexpected outage has caused some other problems that he needs to address before the actual server migration, which has been bumped to next Wednesday...but at least the server is back up and running for now...

 

[Assimilator1]

It is? Ok then I need to bump my cache up again!

 

[Assimilator1]

Why is it that I've now bumped up my cache a 3rd time & BOINC still isn't downloading more WUs! I've set it to get 3days worth + 2days (an increase of 1 day on each) & all it did was d/l 1 extra WU! lol
I've only got 32 WUs! :| (takes ~40mins/WU).

 

[Sunny129]

Fardringle had a similar problem with the LHC@Home project not too long ago. unfortunately he never did figure out a fix...his BOINC client just decided to start fetching new LHC@H work at some point, and we couldn't explain why...

i know you already have some work in the queue that you would lose, but maybe resetting the project might help. once in a while i'll notice my BOINC client downloading a new "master file" from the A@H server prior to fetching new work, which suggests to me that the master file whose presence was necessary in order to complete previous A@H tasks went missing at some point. i've also seen this happen with other DC projects in the past too, and sometimes the communication between my client and those projects' servers is lacking to the point where my client won't download the requisite master file(s). in those cases, the only thing i know to do in order to get my client to fetch the requisite master file(s) is to either reset the project, or detach my client from the project and then reattach.

 

[Assimilator1]

Ok I've just tried a reset, it's now sent me 46 WUs.
So @ ~40mins/WU (=~0.65hr), 46x0.65= 29.9hrs/4 (quad CPU) = 7.475 hrs work! Wth? that's not even the default of 1+1 day!

Computing preferences> Network usage is the right place to set cache levels right?

[update]
Ok I've just detached/reattached A@H, straight away it tried to get 59 WUs , although a bunch failed to d/l!
Still even 59 isn't much, has it got to work out the running average needed?

 

[Sunny129]

Ok I've just tried a reset, it's now sent me 46 WUs.
So @ ~40mins/WU (=~0.65hr), 46x0.65= 29.9hrs/4 (quad CPU) = 7.475 hrs work! Wth? that's not even the default of 1+1 day!

Computing preferences> Network usage is the right place to set cache levels right?

[update]
Ok I've just detached/reattached A@H, straight away it tried to get 59 WUs , although a bunch failed to d/l!
Still even 59 isn't much, has it got to work out the running average needed?

yeah it might be that after resetting a project or detaching & reattaching to a project that BOINC clears all previously known values like average run time, DCF (duration correction factor), etc. and so it may have to learn all of that all over again as you complete more and more A@H tasks. my guess is that your queue will eventually fill up to however many day's worth of work you have it set to right now after you return a few (or several) results and your BOINC client can recalculate the DCF and adjust.

 

[GLeeM]

straight away it tried to get 59 WUs , although a bunch failed to d/l!

Same thing happens to me ... I find if I wait three minutes and then click Update I will get another batch ... then I repeat until I only get one or two. Sometimes I have to click "No new tasks" for the GPU project before Update on the CPU project.

 

[Assimilator1]

I don't know wth is going on with BOINC, last night it ran the cache down ~6 WUs, currently it's got 23 WUs which is just 3 3/4hrs worth!
Any ideas why it's not keeping the cached topped up?

Just updated it & got 60 WUs (only 2 failed to d/l), that's less than 10hrs worth! What's going on?

 

[networkman]

Has anyone posted these questions on the tech forums for their respective projects? I know there isn't much traffic on the Asteroid forum, but it may be others are also seeing the issue.

It also might not be project specific, but BOINC in general?

 

[Sunny129]

on the A@H message boards i see plenty of threads about the project server up and down issues, but i don't see anything specific about not being able to maintain a work queue of specific size. despite all the network problems the server has been having lately, and despite all the download errors as of late, i'm still able to maintain a 2 day queue of A@H work...so i'm not sure at all what is going on w/ Assimilator1's host.

 

[networkman]

BTW, due to some ongoing issues of basically no activity/interest (by the Founder no less), I have switched the "GRRO" moniker back to my "networkman" account on this project and the Einstein@Home project. I am the only member of the aforementioned group that was and has been dedicating computer resources to the DC projects so I feel justified in making this change.

If nothing else, it'll be easier for my DC comrades to track where I'm at and chart progress for races and such.

 

[Sunny129]

...and here i thought that the Grand Rapids Radio Observatory was just your own private home observatory with a professional-sounding name. on the other hand, i don't know of too many folks with a private observatory that operates outside of the visible light spectrum (notwithstanding the use of astronomical filters). glad you're crunching for the TeAm under either moniker...

 

[networkman]

It was intended to be a network of various people with observation posts at each of their houses, where the "observing" was actually happening far beyond visual spectrum and the receivers would then transmit their compiled data to a central location for further computations. But it's hard to consider 2 locations to be much of a group and certainly not enough for any substantial observations to be independently confirmed.

Compounding that issue is the inability of the programmers to move beyond Win95/98 or to more recent hardware advances beyond a Pentium III and you have the makings of a disaster. That's why I'm officially out.

 

[networkman]

There's a new server notice:

"New date of server migration is planned on friday 20th. It should start at about 07:30 UTC and supposed needed time is about 6 - 8 hours."

 

[Assimilator1]

Good to see your name back NWM , but it's better that your with us with any name .

PIII & Win98?? Lol, wow that harks back to the old days! I had a PIII 650 @820 ..... IIRC!

Sunny, NWM et al
What's the lowest you've seen your client cache drop to?

I've got just 19 WUs atm.