Cuda is as Cuda does,,,,,,,,,,,,,,,,,

Message boards : Number crunching : Cuda is as Cuda does,,,,,,,,,,,,,,,,,
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852275 - Posted: 11 Jan 2009, 17:16:31 UTC - in response to Message 852258.  

seti credibility is already going downhill as a result of the mismanaged cuda force-feeding and the deafening silence from the admins on why it was screwed up so badly. Why dont they just come clean and explain exactly what happened in the dark corner of the parking garage where the whole deal went down with a single handshake and the passing of an envelope.

Now you are getting close..........
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852275 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 852296 - Posted: 11 Jan 2009, 17:37:12 UTC - in response to Message 852177.  

Maybe they already cashed the check and had to launch...? Just a thought...


We should never ascribe to malice what is equally explained by fatigue.

The entire staff is something on the order of six people. They do everything from design, through implementation, and on to operations.

If we want more communication from the project, I don't see that happening without staffing someone to do PR -- and if they're going to staff someone, I'd rather see someone who could do actual work, not someone whose job is to paint happy faces on everything.
ID: 852296 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852311 - Posted: 11 Jan 2009, 17:55:26 UTC - in response to Message 852296.  

Maybe they already cashed the check and had to launch...? Just a thought...


We should never ascribe to malice what is equally explained by fatigue.

The entire staff is something on the order of six people. They do everything from design, through implementation, and on to operations.

If we want more communication from the project, I don't see that happening without staffing someone to do PR -- and if they're going to staff someone, I'd rather see someone who could do actual work, not someone whose job is to paint happy faces on everything.

Ain't many happy faces 'round here lately.........I think the Cuda bit their noses off.......
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852311 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852330 - Posted: 11 Jan 2009, 18:30:10 UTC - in response to Message 852329.  
Last modified: 11 Jan 2009, 18:32:08 UTC

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Ooooooops....I think a hitman from nvidia is knocking at my door......
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852330 · Report as offensive
Profile SATAN
Avatar

Send message
Joined: 27 Aug 06
Posts: 835
Credit: 2,129,006
RAC: 0
United Kingdom
Message 852339 - Posted: 11 Jan 2009, 18:41:17 UTC

knock knock.

Oi Kittyman!
ID: 852339 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 852340 - Posted: 11 Jan 2009, 18:41:44 UTC - in response to Message 852330.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Ooooooops....I think a hitman from nvidia is knocking at my door......

Bargain with him and get some info on new cards that are coming in the next 6-12 months!
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 852340 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852342 - Posted: 11 Jan 2009, 18:42:36 UTC - in response to Message 852339.  

knock knock.

Oi Kittyman!

Got my Browning at the ready.......LOL.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852342 · Report as offensive
Beau

Send message
Joined: 24 Feb 08
Posts: 50
Credit: 129,080
RAC: 0
United States
Message 852343 - Posted: 11 Jan 2009, 18:42:58 UTC - in response to Message 852339.  

Send out the attack cats!!!
ID: 852343 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 64940
Credit: 55,293,173
RAC: 49
United States
Message 852344 - Posted: 11 Jan 2009, 18:44:13 UTC - in response to Message 852342.  

knock knock.

Oi Kittyman!

Got my Browning at the ready.......LOL.

Yeah You better have the BAR well maintained and ready for Bear. ;)
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 852344 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 852348 - Posted: 11 Jan 2009, 18:45:28 UTC - in response to Message 852330.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Actually, Nvidia stands to lose big-time if CUDA fails.

I haven't looked at the code, but it appears that the only real change is to call the CUDA FFT library instead of the FFTW library, the rest of the client is unchanged.

That's the same kind of change that the optimizers did to replace FFTW with the (faster, licensed) Intel math library.

If it produces slightly different numbers that's understandable -- we aren't talking about something that emulates the 8087 FPU, and FFTs do so many math operations that any tiny error is likely to accumulate. That happens even between different generations of Intel and AMD chips.

However, if the CUDA FFT code is a "black box" and it's the black box that crashes, that means the Nvidia capability is worthless.

More to the point, it means their expensive, fancy "Tesla" supercomputer can't do math.

ATI is looking fairly clever right now.
ID: 852348 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852350 - Posted: 11 Jan 2009, 18:47:28 UTC - in response to Message 852348.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Actually, Nvidia stands to lose big-time if CUDA fails.

I haven't looked at the code, but it appears that the only real change is to call the CUDA FFT library instead of the FFTW library, the rest of the client is unchanged.

That's the same kind of change that the optimizers did to replace FFTW with the (faster, licensed) Intel math library.

If it produces slightly different numbers that's understandable -- we aren't talking about something that emulates the 8087 FPU, and FFTs do so many math operations that any tiny error is likely to accumulate. That happens even between different generations of Intel and AMD chips.

However, if the CUDA FFT code is a "black box" and it's the black box that crashes, that means the Nvidia capability is worthless.

More to the point, it means their expensive, fancy "Tesla" supercomputer can't do math.

ATI is looking fairly clever right now.

nvidia is already smiling.......they no doubt have sold a bunch of high powered vid cards in the wake of this fiasco.......and that is just what they wanted out of it.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852350 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 64940
Credit: 55,293,173
RAC: 49
United States
Message 852355 - Posted: 11 Jan 2009, 18:52:13 UTC - in response to Message 852350.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Actually, Nvidia stands to lose big-time if CUDA fails.

I haven't looked at the code, but it appears that the only real change is to call the CUDA FFT library instead of the FFTW library, the rest of the client is unchanged.

That's the same kind of change that the optimizers did to replace FFTW with the (faster, licensed) Intel math library.

If it produces slightly different numbers that's understandable -- we aren't talking about something that emulates the 8087 FPU, and FFTs do so many math operations that any tiny error is likely to accumulate. That happens even between different generations of Intel and AMD chips.

However, if the CUDA FFT code is a "black box" and it's the black box that crashes, that means the Nvidia capability is worthless.

More to the point, it means their expensive, fancy "Tesla" supercomputer can't do math.

ATI is looking fairly clever right now.

nvidia is already smiling.......they no doubt have sold a bunch of high powered vid cards in the wake of this fiasco.......and that is just what they wanted out of it.

And now they've brought their GTX 295 card which is essentially two GTX 260 cards stuck together and die shrunk to 55nm from 65nm all while using less power too from what I've read online. While ATi's 4870 X2 card uses more power like It's going out of style and puts out more heat to boot. :o
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 852355 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 852370 - Posted: 11 Jan 2009, 19:14:00 UTC - in response to Message 852275.  

seti credibility is already going downhill as a result of the mismanaged cuda force-feeding and the deafening silence from the admins on why it was screwed up so badly. Why dont they just come clean and explain exactly what happened in the dark corner of the parking garage where the whole deal went down with a single handshake and the passing of an envelope.

Now you are getting close..........


What if the contents of the envelope help to get a fat pipe up the hill???
Wouldn't that be a good thing??
Or keep the project online for another year??
Also a good thing!


Boinc....Boinc....Boinc....Boinc....
ID: 852370 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 852374 - Posted: 11 Jan 2009, 19:17:39 UTC - in response to Message 852350.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Actually, Nvidia stands to lose big-time if CUDA fails.

I haven't looked at the code, but it appears that the only real change is to call the CUDA FFT library instead of the FFTW library, the rest of the client is unchanged.

That's the same kind of change that the optimizers did to replace FFTW with the (faster, licensed) Intel math library.

If it produces slightly different numbers that's understandable -- we aren't talking about something that emulates the 8087 FPU, and FFTs do so many math operations that any tiny error is likely to accumulate. That happens even between different generations of Intel and AMD chips.

However, if the CUDA FFT code is a "black box" and it's the black box that crashes, that means the Nvidia capability is worthless.

More to the point, it means their expensive, fancy "Tesla" supercomputer can't do math.

ATI is looking fairly clever right now.

nvidia is already smiling.......they no doubt have sold a bunch of high powered vid cards in the wake of this fiasco.......and that is just what they wanted out of it.

I don't think it's that simple.

If they move a bunch of cards now, and it doesn't work, they've mortgaged their future against today's failure.

That doesn't build brand loyalty or repeat customers.
ID: 852374 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852377 - Posted: 11 Jan 2009, 19:19:43 UTC - in response to Message 852374.  

You mean nvidia wouldn't provide someone part time to do pr work for them? They should have made that clear before they tucked the envelope into their trench coats and walked away grinning....


Oh......nvidia got what they wanted.....and left Seti holding the bag......

Actually, Nvidia stands to lose big-time if CUDA fails.

I haven't looked at the code, but it appears that the only real change is to call the CUDA FFT library instead of the FFTW library, the rest of the client is unchanged.

That's the same kind of change that the optimizers did to replace FFTW with the (faster, licensed) Intel math library.

If it produces slightly different numbers that's understandable -- we aren't talking about something that emulates the 8087 FPU, and FFTs do so many math operations that any tiny error is likely to accumulate. That happens even between different generations of Intel and AMD chips.

However, if the CUDA FFT code is a "black box" and it's the black box that crashes, that means the Nvidia capability is worthless.

More to the point, it means their expensive, fancy "Tesla" supercomputer can't do math.

ATI is looking fairly clever right now.

nvidia is already smiling.......they no doubt have sold a bunch of high powered vid cards in the wake of this fiasco.......and that is just what they wanted out of it.

I don't think it's that simple.

If they move a bunch of cards now, and it doesn't work, they've mortgaged their future against today's failure.

That doesn't build brand loyalty or repeat customers.
And this does not surprise you?

"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852377 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852379 - Posted: 11 Jan 2009, 19:22:34 UTC - in response to Message 852378.  


What if the contents of the envelope help to get a fat pipe up the hill???
Wouldn't that be a good thing??
Or keep the project online for another year??
Also a good thing!



You're assuming that the envelope was directed at SAH and that's where your missing DA in that equation. DA is the puppet master who pulls the strings attached to his lab rat (SAH) ;)

Why else would NV post a pic of DA on their press release ?




Oh........now you are getting sooooooooooooo close....

"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852379 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9949
Credit: 103,452,613
RAC: 328
United Kingdom
Message 852381 - Posted: 11 Jan 2009, 19:23:37 UTC

Whilst I am no SETI@Home CUDA supporter, I wonder exactly how many cards extra high end cards Nvidia have sold. In the grand scheme of things I wouldn't have thought that many, and as several people have already said if CUDA turns out not to work. it will lok pretty bad for Nvidia and I for one will be looking to ATI.

Bernie.
ID: 852381 · Report as offensive
Beau

Send message
Joined: 24 Feb 08
Posts: 50
Credit: 129,080
RAC: 0
United States
Message 852404 - Posted: 11 Jan 2009, 20:10:59 UTC - in response to Message 852389.  

I'll just take that as being I hit the nail on the head.
ID: 852404 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51464
Credit: 1,018,363,574
RAC: 1,004
United States
Message 852405 - Posted: 11 Jan 2009, 20:16:38 UTC - in response to Message 852404.  

I'll just take that as being I hit the nail on the head.

Ya typhooned it.......
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 852405 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 852438 - Posted: 11 Jan 2009, 21:00:30 UTC - in response to Message 852427.  

Better ways indeed, especially if this whole CUDA thing blows up in Nvidia's face.


PROUD MEMBER OF Team Starfire World BOINC
ID: 852438 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : Cuda is as Cuda does,,,,,,,,,,,,,,,,,


 
©2022 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.