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

[Assimilator1]

Yea I was away this weekend, came back just now to find the client had ran dry .
It started to report no WUs available from 12.53am here & then ran out at ~2am

Hmm, I'd like to do a uninstall/re-install to sort out this stupid cache issue now but I just managed to get 15 WUs which I may not be able to get again......

 

[Sunny129]

well according to the project server status page, there's still hardly any new work to go out (this despite the fact that i've managed to DL almost 100 new SSE3 tasks since this morning when i had 0 tasks), so now might be a good time to set NNt, finish the 15 new tasks in your queue, and then do the uninstall/reinstall. you're probably going to want to do it sooner or later anyways...

 

[Assimilator1]

OK I've un/re-installed it, the initial WU grab was 62, I think that's more then it's done before, I'll see how it goes once it's worked out the runtime average.

 

[Assimilator1]

OMG I'm losing it with BOINC, it's running the cache down to just 26 WUs (~10hrs), just hit update & it's still only got 52 WUs, wth!?

Now should I hit the BOINC forums or A@H about this cache problem?

 

[Sunny129]

well, if you're having the same cache issues with other projects as well, then i'd hit the BOINC message boards. if you're not sure, you'll unfortunately have to attach to another project (preferably a stable one that constantly has new work to be sent out) and see what the work fetch behavior is for that project. i know its time consuming, but i don't know how else to approach the problem. aside from reinstalling BOINC, have you tried detaching and reattaching to the Asteroids@Home project?

 

[Assimilator1]

Only running A@H so don't know about other projects.

Already un/re-installed BOINC, & yea have previously tried detaching project.
I'm not going to try another project to figure out the cache problem, too time consuming + lost crunching time for A@H.
I'll hit the A@H forum & see how it goes.

 

[GLeeM]

Usually A@H will not fill up the queue with one try - I will get ~50 one time and then ~70 or ~40 the next update.

 

[Assimilator1]

I can't believe I've missed this, I've just found in the network usage tab the cache limits! :$, previously I had been altering the BOINC cache limits through the online preferences (to no effect), which makes me wonder, what's the point in it if where the client takes priority?

Anyway I've now got 449 WUs cached, which works out to 3 days worth (as per my new setting).

......... still can't believe I missed the network tab settings! :$, although I'm not the only one whose missed that here .

 

[Sunny129]

......... still can't believe I missed the network tab settings! :embarrassed:, although I'm not the only one whose missed that here .

i don't think any of us would have suspected that your actual client cache settings were the problem LOL. in fact, short of rereading the entire thread to find any mention of your cache settings, if you had made such a mention, i think all of us probably assumed that you were talking about the cache settings accessible through the BOINC manager, and not the web preferences, thus explaining how it got so easily overlooked. at any rate, i'm glad that after several weeks of cache issues, you've finally figured it out. more often than not the problem is silly, the solution simple, and usually its sitting right under our noses haha.

 

[Assimilator1]

Lol, sure was this time , I did a quick scan of the previous page earlier & I did ask if I was looking at the right options in network usage but I didn't state whether it was from the client or the web preferences & it wasn't clear which it was , didn't scan back any further, oh well, nm lol.

So what is the point in the web preferences then? (seeing as the caches settings there have no affect).

 

[Sunny129]

well there are 3 categories of web preferences - computing preferences, community preferences, and Asteroids@Home preferences. the settings that are supposed to affect how the BOINC client behaves on your local machine, but appear not to work are the "computing preferences." i'm not 100% certain that these settings don't do anything so much as i am certain that the computing preferences available in your BOINC manager override the web "computing preferences." the web "community preferences" do in fact function, and they simply control your online community/message board preferences. and finally there is the "Asteroids@Home preferences" category - these preferences are also functional, and they control such things as how much resources you want BOINC to allocate to A@H project, or which application (or applications) you'd like to run and which ones you don't want to run. with A@H its a bit oversimplified b/c you don't really have choices here (there is only one A@H application depending on your platform). on a project that runs multiple applications, the "________@Home preferences" would be where you select which ones you want to run.

 

[Assimilator1]

Yea I know the forum preferences do work, I was referring to the BOINC web preferences.
I just don't get what they're for seeing as the client settings take precedence, even with BOINC on it's default settings. I could understand it if the web options worked until you change from the default, but they don't, wierd

 

[networkman]

So we [the Asteroids@Home project] are apparently down for some kind of "maintenance" again. No notices or updates AFAIK. Hopefully, you all adjusted your caches adequately. :/

 

[Sunny129]

there is a "sudden maintenance" notice on the A@H home page...of course it wasn't given to us with much advance warning if any at all. Kyong doesn't elaborate beyond that, so i'm assuming that by sudden, he means "unexpected." hopefully we'll be back up and running soon...

 

[Assimilator1]

Really? Maybe it's sorted now then because my cache is full & no warning messages in the log currently....... ah yea looking back I see it couldn't get any WUs from ~12.50-15.50.

Cache level all sorted now so no probs

NWM
I just can't quite close the production gap on you!
It would help if I didn't have to shut a core off sometimes to keep temps down .

 

[networkman]

Assim: If it helps, I did have to shut down one of my boxes (as it was just a stress-test machine), so you should start to see some drop in my production that you may be to take advantage of.. that or I'm just teasing you.

 

[Assimilator1]

Lol, well as it happens I've also been able to keep all 4 cores crunching so my RAC has gone up considerably & I'm now outgunning you! .

I've also upgraded the front fan on my rig (about to do the rear too) so they'll be less time powering it down .

 

[GLeeM]

The WUs have almost doubled in length AND the points have doubled

 

[Assimilator1]

Oh yea! I haven't got to the long WUs yet, still got plenty of short ones to crunch through .
I noticed WUs with a deadline of Oct 25th are a mixture of short & long WUs.

 

[Assimilator1]

Wth? Recently my rig did a WU in 910s, the 'short' ones take ~2400s, so what's this iddy biddy WU? lol

Btw it was validated so it's not an errored WU, see here http://asteroidsathome.net/boinc/workunit.php?wuid=8624105

Anyone else had these?

 

[Sunny129]

yes, i've had a few...tough the run times were approx. 50% that of a normal WU (compared to your ~38%), and i only got half credit for them (120)! i see you got full credit for your shorty! wth indeed LOL

 

[Assimilator1]

Oh yea! I hadn't noticed the credit! Stranger & stranger said Alice .

Btw, don't mean to curse anything, but it seems the servers have been reliable lately .

 

[Sunny129]

true...99% of my completed tasks are validating. right now i'm fortunate enough to see less than one invalid or error per day.

 

[Assimilator1]

Cool .

Not crunching just atm, upgraded the CPU from the Q6600 to Q9550, doing stability testing on it atm, currently o/ced to 3.55 GHz. I think the CPU's got a lot of headroom left , not sure about the mbrd though.

 

[Sunny129]

i did get a bunch of errors the other day (hundreds of them), but they were of the "timed out- no response" type, so i think it was an issue with BOINC miscalculating my hardware's capabilities and the associated work buffer, as opposed to more server download issues...