http://www.nvidia.com/object/cuda_home.html

"
SETI@home
SETI (Search for Extraterrestrial Intelligence) is a scientific area whose goal is to detect intelligent life outside Earth. One approach, known as radio SETI, uses radio telescopes to listen for narrow-bandwidth radio signals from space. Such signals are not known to occur naturally, so a detection would provide evidence of extraterrestrial technology. You can participate by running this free program that downloads and analyzes radio telescope data. And now, with this CUDA-optimized client powered by your CUDA-ready GeForce GPU, your system can deliver as much as 10 times the computational power of an average home PC CPU.
Author(s): University of Berlkey Application Type: Signal Processing
Organization Type: Consumer Speed Up: 10 x
Organization: University of Berlkey Date Released: December 2008
Software License: Available Content: Application
"

I think it's very good for SETI project and such announce will not only make nVidia some sell boost but will bring SETI@home many new participants also.
Consider it as free advertisment on some pretty populated street.

I think they should have stayed around for long enough to cure the VLAR bug before making a statement like that. We've seen too many people come here, give it a try, find their desktop unusable, and abandon both SETI and CUDA - to the detriment of both endeavours. That's the bad sort of free advertising.

Yes, I know there are kludges and workrounds - and full kudos to you, Raistmer, for making them possible - but what the people who read the nVidia blurb need is a stock Berkeley application that works straight off the server, on all ARs.

but what the people who read the nVidia blurb need is a stock Berkeley application that works straight off the server, on all ARs.

Yes, in some aspects CUDA not for "faint hearts" :D
But it could rise interest in some of peoples who just never heard about SETI, let them think about concept and so on and so forth. Some can refuse do CUDA part but join as CPU-only members, some can mention about SETI in conversation with friend and that friend will take additional effort to run app and so on.
We, SETI participants, tend to think all know about SETI but just not all wanna participate in it... But reality is far far away from this. In general, very little amount knows about SETI at all... :(

But reality is far far away from this. In general, very little amount knows about SETI at all... :(

That's so true!

I was surprised there are no iGoogle Themes for SETI or nVidia's CUDA. Don't we have any graphics artists around here??

Martin

but what the people who read the nVidia blurb need is a stock Berkeley application that works straight off the server, on all ARs.

Yes, in some aspects CUDA not for "faint hearts" :D
But it could rise interest in some of peoples who just never heard about SETI, let them think about concept and so on and so forth. Some can refuse do CUDA part but join as CPU-only members, some can mention about SETI in conversation with friend and that friend will take additional effort to run app and so on.
We, SETI participants, tend to think all know about SETI but just not all wanna participate in it... But reality is far far away from this. In general, very little amount knows about SETI at all... :(

Just to give a different perspective on this; although I've read up on the optimized versions, and I'm familiar with the VLAR problem, I simply opted to leave everything as stock. From someone who's only picked SETI back up this past October, things seem to be working reasonably well. When I find a VLAR rumbling through, I simply shrug and let it run to completion.

Wasted cycles? On occasion.

Does it matter in the greater scheme of things? Only if you want it to.

I suspect some newbies will be put out on that rare occasion that they've come to the project, panting for an instantly high RAC. If they do, they're likely not coming to the table with P3's and 8600 cards. A quad is likely more the case with the latest CUDA-capable offering from the marketplace tucked inside. That will do away with the occasional VLAR on the stock app well enough, whilst fast moving MB WUs flicker by and satisfy the desire for "Swifter, Higher, Stronger."

A new card running CUDA has been a great boon to one of my slow machines. The others had nVidia cards that were getting long in the tooth anyways and CUDA didn't give much of a lift to them regardless.

If nothing else, I would put my vote behind getting a CUDA-enabled AP app (stock please as it otherwise means extended fiddling that limits its potential adopters).

Hope the above is a useful view and not too far off the mark.

Regards,

---

Unfortunately they are using the same 10x marketing scam... There is no 10x speed up, more like another core added when we are comparing equal amounts of money spend.

Unfortunately they are using the same 10x marketing scam... There is no 10x speed up, more like another core added when we are comparing equal amounts of money spend.

They say "as much as 10 times the computational power of an average home PC CPU". I don't think that's far off, even considering price since most PCs are bought as brand name preconfigured systems.

Granted, the impression from a quick reading may cause some to overestimate the GPU capabilities, but most will have a reasonable amount of skepticism.

                                                               Joe

True, that statement is vague enough that it could be eventually true. Like pinning the cheapest CPU against the most expensive CUDA card.
But that kind of verbage it is characteristic usually for deceiving marketing. It is sad that an scientific endeavour got mixed-up in such advertising.
Why? Because some of the new commers might feel that where scammed or someting and a negative effect of this will affect Seti credibility as a whole.
I guess it was a part of the deal... I don't know.

And it doesn't matter to much when "no work is available for SETI@home Enhanced"...

I think most new folks will be the plug it in and lets try seti types. And i agree that when id wont work right the will be storm clouds. It would be nice if before you could down load seti that you had to peruse the message boards and take a test:) Thats why i think the check boxes should be in the no Gpu and no AP untill the user knows whats up. And even though Ive been crunching for almost 10 years I didnt even come to the message boards for 9 and a half years. What a waste of knowledge I didnt get . I wont do opp apts yet because i dont think i can, at least not the first time by myself. Just my 2 cents worth.

---

[/quote]

Old James

BTW.

AMD Phenom II X4 940 BE @ 4 x 3.0 GHz [AK v8.0 SSE3]
nVIDIA Geforce GTX260 Core216 [55nm] (OC-Edition)

A MB AR-WU=0.44x:
On the GPU [~ 8 min.] ~ 7.5 x faster as on CPU-Core [~ 60 min.].

---

Is it true that a 295 card runs close to 300 watts? My i7 machine runs at less than 300 watts. So a 295 and my i7 would run at less than 600 watts if that true is too much juice to buy for me.