I'm not getting any cuda work today....

I know this happens sometime, just wanted to see if others agree.

I fried like 200 cudas doing some GPU testing yesterday....any chance I'm being punished with no work?

There was no problem with the scheduler or downloading today that I know of, but looks like your 560 box just got about 50 GPU WU's, so you should be good now. But yes, your task limits WILL suffer from returning too many errors. They will also increase with consecutive validated units.

---

Okay...scheduler issue today? (Is there a place I can check this, I'm sure...I'll google..)

lol My guess is YES...because some people got their hands on the GTX 680 and started guzzling the cudas down.

Mine won't be here until March 30th. And when the call it the Kepler series...they are talking to us, si? sweet,

Okay...scheduler issue today? (Is there a place I can check this, I'm sure...I'll google..)

lol My guess is YES...because some people got their hands on the GTX 680 and started guzzling the cudas down.

Mine won't be here until March 30th. And when the call it the Kepler series...they are talking to us, si? sweet,

The place to check if others are having problems is in the Number Crunching Forum, in a thread called "Panic Mode On". Also, watch the News Forum, and occasionally the Technical News Forum.

But the short answer is yes, there was a problem wednesday night, it was corrected Thursday morning, and everyone is trying to get a slice of the pie now. This will last awhile.

No one knows exactly how effective the kepler series GPU's will be on this project. It's all new architecture, and therefore, could either be great, or not so much. The first release (680) seems more aimed at gamers. Later releases may be more CUDA oriented.

---

Thanks!

Still no cudas for me :(

We need more telescopes? Or more WU makers? Or is this ALLL server-side issues?

More a congestion of the connection out of Seti. The green is their uploads to us, as you can see it's maxed out.

So essentially, what this project needs is a better connection to the internet. Now, the lab has a 1Gbit connection, but Seti is only allowed to use 100Mbit of that. University of California - Berkeley politics. Nothing we can do about that.

---

i c i c


What would it take for us to get seti a dedicated 1Gbit? Humor me if need be...

You'd have to fight the politics at the University of Berkeley, really. Or perhaps if some rich person said, here's a million, go get it fixed, that that would do it. But even he would have to fight the politics at the University, as the lab where Seti runs from is a university building, the ground that the cable would need to run through, the whole mile down the hill, is university grounds.

---

lol, how about a couple T3s supported by the collective??

You would need at least 2 T3 lines (44.6Mbps) to equal what they're using now, which is probably more expensive than what they are using now, and you'd still have to play with Berkeley politics to get the line run down to the building they're in.

...and what would happen to those T3 lines if the "collective" simply couldn't donate enough per month or per year to keep the lines going?

I would image that based upon the crunching power donated to SETI, they would likely need something beefier like an OC-768 connection (38,486Mbps) to last them a while... anyone have that million Ageless requested ealier? ;)

Okay...I'm 10minutes from Berkeley...can I hardwire in?

hahahahah only half kidding.