The first task which I think failed was Task 2363542721

The second which seemed to run ok was 2363542731

At a normal setting for the gpu the task used was 2363542787

Rather aggressive overclock was task 2363542799

Slightly less aggressive overclock was task 2363542797

Looks like 2 of the 4 valid tasks on your machine are from the 680. With the only failures being download errors. So lets hop the rest look as good!

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

"Multibeam x41g Preview, Cuda 3.20"

That's a relief - and thanks Hal for the host link.

Please, other testers/early adopters - don't run with computers hidden and an aggressive overclock, until we know that the old software runs on the new hardware.

Edit - both download errors were "user requested transfer abort". Why was that, may I ask?

I've not aborted any tasks for last week or so - only aborted those as they seemed to get stuck downloading.

Got another 6 tasks now stuck trying to download which is frustrating.

As a side note it looks like the 680 does not work on the GPUGRID project - downloads the work units ok then crashes out when it starts with acemd.win2352 error.

---

I've not aborted any tasks for last week or so - only aborted those as they seemed to get stuck downloading.

Got another 6 tasks now stuck trying to download which is frustrating.

As a side note it looks like the 680 does not work on the GPUGRID project - downloads the work units ok then crashes out when it starts with acemd.win2352 error.

There is a retry button for downloads.

@MartinBen:

Do tell, are you running multiple (how many) WUs concurrently?

What is the GPU-loading?

Thanks.

There is a retry button for downloads.

Yes I know that!

But just how many times do I have to keep hitting "retry" before the server actually pushes the work out?

---

@MartinBen:

Do tell, are you running multiple (how many) WUs concurrently?

Thanks.

Let's start with 1, shall we, until we know it works with SETI software? Speed comparisons can come next.

I've not aborted any tasks for last week or so - only aborted those as they seemed to get stuck downloading.

Got another 6 tasks now stuck trying to download which is frustrating.

As a side note it looks like the 680 does not work on the GPUGRID project - downloads the work units ok then crashes out when it starts with acemd.win2352 error.

There is a retry button for downloads.

or try suspend comms then reset availability about 10 seconds later.

@MartinBen:

Do tell, are you running multiple (how many) WUs concurrently?

Thanks.

Nope only 1 x GPU work unit at a time - that is assuming we can actually download them to start with which seems to be a constant issue with this project.

---

Thanks! 1x WU at a time is best.

What's the GPU-loading?

Thanks again.


EDIT: Geezzz... SAH SMS board at work! :D

Martin has one inconclusive workunit so far: WU 955095540

This is the sort of warning sign testers will be looking out for, but in this case I'm inclined to suspect that the wingmate is at fault - he has a number of other invalid results.

In theory the 680 should be twice as good @seti than the 580...

Well so much for theory!

I'll get me coat:)

Yes leaving the card slightly overclocked seems to be fine for doing WU's.

On GPUGRID - they have confirmed that project does not currently support the 680 which is a shame but they hope to have a fix very shortly.

Einstein - Strange as can't get any Gravitational Wave WU's to run but the Gamma-ray pulse WU's seems to be running just fine.

---

Hi,
Would you like 15 or 20?:-)
I got overloaded with the damm things and they are taking forever on my rig:-/ Only got GTX560's and they are taking over 9 hours..
Preempting everything else right now.

Regards,

---

Cliff,
Been there, Done that, Still no damm T shirt!

On GPUGRID - they have confirmed that project does not currently support the 680 which is a shame but they hope to have a fix very shortly.

Well, skgiven in your thread said it wasn't compatible yet, but he's a moderator and tester, rather than a programmer or scientist. I don't know whether the hint I got from your errored-out tasks (GPUGrid may be struggling with the 2 GB VRAM detection) is correct or not - but from the news threads, somebody donated them a card for testing this morning, so as you say it shouldn't take long to track down whatever it is.

Yes leaving the card slightly overclocked seems to be fine for doing WU's.

On GPUGRID - they have confirmed that project does not currently support the 680 which is a shame but they hope to have a fix very shortly.

Einstein - Strange as can't get any Gravitational Wave WU's to run but the Gamma-ray pulse WU's seems to be running just fine.

Youre' not by chance getting the message from the server that your driver version is too large?

---

I'm not the Pope. I don't speak Ex Cathedra!

Einstein - Strange as can't get any Gravitational Wave WU's to run but the Gamma-ray pulse WU's seems to be running just fine.

Those are CPU only Wu's that are erroring out, you haven't done any BRP4cuda32 Wu's at Einstein on your 680 yet.

All tasks for computer 3740033

Claggy

Edit: fixed link.

You're not by chance getting the message from the server that your driver version is too large?

The driver version check is on the Binary Radio Pulsar Search apps only. Gravity Wave searching hasn't been ported to CUDA - it's CPU only.

Here's a website where their doing a 4 way GTX680 review...

Oh and it's a Dutch site and the review is in English, nice...

That looks like a Silverstone ST1500 1500w psu too, neat.

It'll be a miracle if it works in any game. There are so many problems with quad-SLI from what I gear.

Here's a website where their doing a 4 way GTX680 review...

Oh and it's a Dutch site and the review is in English, nice...

That looks like a Silverstone ST1500 1500w psu too, neat.

Meowwrrrrrr!!!
Looks like a pretty tough piece of kit.
Kitties want!

---

"Time is simply the mechanism that keeps everything from happening all at once."