Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

Check if your wingman also reports it as -9 overflow. If that's the case it is likely to be a real -9, especially if your wingman crunch it on his CPU instead of his GPU.

Sten-Arne

Okay, thank you. Hopefully the servers will be fixed soon. I want to start crunching again. :(

No problems.

If you up your cache to at least a couple of days, you will not be vurnerable to server down time. You can still crunch even if the servers aren't available, using cached work units.

(In Boinc Manager: Advanced/Preferences, network usage tab, and setting the "additional work buffer" to at least a couple of days). When the servers starts delivering again, you will receive additional work units.

Sten-Arne

Yeah, i've now set my cache to 7 days up from 3, i haven't gotten many new WU's in the last 3 days. :P

Great, now we just have to wait for things to get back to normal again. I don't expect the servers to start delivering new WU's until after the weekly normal outage which will happen today (heh, we will have an outage during an outage :-).

At least the message board is alive.....

Sten-Arne

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

Check if your wingman also reports it as -9 overflow. If that's the case it is likely to be a real -9, especially if your wingman crunch it on his CPU instead of his GPU.

Sten-Arne

Okay, thank you. Hopefully the servers will be fixed soon. I want to start crunching again. :(

No problems.

If you up your cache to at least a couple of days, you will not be vurnerable to server down time. You can still crunch even if the servers aren't available, using cached work units.

(In Boinc Manager: Advanced/Preferences, network usage tab, and setting the "additional work buffer" to at least a couple of days). When the servers starts delivering again, you will receive additional work units.

Sten-Arne

Yeah, i've now set my cache to 7 days up from 3, i haven't gotten many new WU's in the last 3 days. :P

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

Check if your wingman also reports it as -9 overflow. If that's the case it is likely to be a real -9, especially if your wingman crunch it on his CPU instead of his GPU.

Sten-Arne

Okay, thank you. Hopefully the servers will be fixed soon. I want to start crunching again. :(

No problems.

If you up your cache to at least a couple of days, you will not be vurnerable to server down time. You can still crunch even if the servers aren't available, using cached work units.

(In Boinc Manager: Advanced/Preferences, network usage tab, and setting the "additional work buffer" to at least a couple of days). When the servers starts delivering again, you will receive additional work units.

Sten-Arne

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

Check if your wingman also reports it as -9 overflow. If that's the case it is likely to be a real -9, especially if your wingman crunch it on his CPU instead of his GPU.

Sten-Arne

Okay, thank you. Hopefully the servers will be fixed soon. I want to start crunching again. :(

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

Check if your wingman also reports it as -9 overflow. If that's the case it is likely to be a real -9, especially if your wingman crunch it on his CPU instead of his GPU.

Sten-Arne

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Is there a way to check if it's a real -9 overflow? Sorry about all the questions, i'm just not very use to all this stuff yet. :P

The problems is only about the excessive overflows of Fermi GPU's

Although it doesn't actually state its the fermis that are causing the problem. It just says GPUs. Perhaps I'm being too paranoid?

No it doesn't state that, but I know that it's about the Fermi's. Browsing around in the "Number crunching" part of this Message board, will soon reveal that Fermi's are the villain.

Sten-Arne

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

Well, they may not be erroneous. They may actually be real -9 overflows. If your wingman also reports them as -9 overflows (especially if your wingman crunch them on his CPU) they are probably real -9 overflows.

Not all -9's are erroneous, there are tons of real -9's too, (The number of results detected exceeds the storage space allocated)

Sten-Arne

Ah okay. Thanks for that, yeah it seems like i've processed 20 or so cuda WU's and only 2 of them were erroneous.

The problems is only about the excessive overflows of Fermi GPU's

Although it doesn't actually state its the fermis that are causing the problem. It just says GPUs. Perhaps I'm being too paranoid?

---

Hey, just wanted to check if this may be related to the overflow issue they're talking about on the main page with GPU WU's

I have 2 WU's that haven't validated yet both have:

"SETI@Home Informational message -9 result_overflow
NOTE: The number of results detected exceeds the storage space allocated."

This bad? I'm kind of new to the whole thing here, i've been around a while but never really understood a great deal of it.

The problems is only about the excessive overflows of Fermi GPU's. Since you have a GTX 275 GPU, it isn't about your GPU.

Most CUDA GPU's seems to produce some erroneous overflows from time to time, but this issue is about Fermi's that does nothing than produce -9 overflows by the thousands.

Sten-Arne

Quite possibly. There doesn't seem to be any more information other than what is posted on the home page. If I were getting thee errors, I would shut down processing on my GPU until the project came out with more definite information or a fix.

---

Hey, just wanted to check if this may be related to the overflow issue they're talking about on the main page with GPU WU's

I have 2 WU's that haven't validated yet both have:

"SETI@Home Informational message -9 result_overflow
NOTE: The number of results detected exceeds the storage space allocated."

This bad? I'm kind of new to the whole thing here, i've been around a while but never really understood a great deal of it.