SETI MB CUDA for Linux

[sunu]

I still can't understand why your out of memory tasks don't fall back to cpu computing.

[riofl]

maybe they can't find room to work? i run an average of 6.0 average load with 420 average number of tasks.

[riofl]

ugh pending is up to 109k!  why does it go so high? is it that they question the results and are waiting verification with someone else? or is it just that i am pushing out higher volume than they expect from me so they suspect most of it?


my machine's average credit dropped from 18.6k down to 16.9k

im thinking maybe it was better at 3.3ghz?

[sunu]

No, hang in there. My pc lost 4000 RAC since yesterday. It's because of Berkeley's connection problems that started yesterday.

[riofl]

ahh ok... i hate it when things happen outside of my control especially when they happen right when i am wanting to compare results.  ok ill let it sit. machine is happy so i can wait.

[riofl]

ugh. looks like its gonna be days before things get back to normal. a/c failure in the server closet.

[riofl]

hehe major changes... their website finally came up but its different. guess they dont have the accounts section done yet either. oh well.. at least boinc works

[Claggy]

Doesn't work in I.E very well, Firefox O,K, I'm wondering if it's supposed to be an Internal Site.

Claggy

[Richard Haselgrove]

It's the old-ish "All SETI at Berkeley" site - much more general than just the '@home' sub-project. Check the links for SERENDIP, SEVENDIP, CASPER etc. across the top. The weekly blogs are often fun too.

I guess they hooked it up to look a bit more interesting than yesterday's placeholder - shame about IE, though. (And actually, the placeholder was more informative for boinc@home users)

I hope the change doesn't mean the main SETI@home links will be offline for an extended period - that Googlehack may have been more serious than Eric thought at first.

[riofl]

yeah i can see that. an old version of the site is better than no version. i just hope they can restore it all. i can just see us losing our accounts.. ugh..

[Claggy]

It's already up.

Claggy

[riofl]

cool thanks! ill go check things out

[riofl]

i am seeing quite a lot of these errors today. may be the tesla

Sun 28 Feb 2010 11:04:58 AM EST   SETI@home   [error] 09mr07ad.12026.32399.16.10.74_1: negative FLOPs left -966550029782.363525

[sunu]

Haven't seen something similar before. Is it something consistent or you've seen it a few times and that's it?

[riofl]

it was quite often during that particular time boinc ran. it is the only time i have ever seen such an error.. i shut the machine down once i noticed that and allowed a 15min cooldown period and then restarted it and everything has been fine since. i think next month i am going to replace the tesla. it appears to be acting up more and more and in strange ways. at least i am assuming it is the tesla. i will have to research this unit below to tell which device it came from.