GF100 Fermi card at GTC a fake / mock-up

[imported_Shaq]

It's damaging towards Nvidia so its gotta be news!!

 

[OCGuy]

As long as Newegg doesnt try that with me, i'll be OK.

 

[thilanliyan]

Did they show the card running any GPGPU code at GTC?

 

[nitromullet]

Originally posted by: thilanliyan
Did they show the card running any GPGPU code at GTC?

I doubt there would be any speculation about this if they did.

 

[Idontcare]

Originally posted by: nitromullet

Originally posted by: Idontcare
But if it makes you feel better to surmise that NV bandied about a "for display purposes only" tesla shell then be my guest, sure it is a fake. Now that we concluded and agreed to that, what does it change? About tesla? About GTX380? About anything?

Well, it might be relevant if it's an indicator of whether or not NV has working cards or not (Tesla or otherwise). Although, obviously even if NV showed off only a non-functional shell, it doesn't necessarily mean that they don't have cards working in their labs.

That would be my concern...we have to travel waaaay down the assumption train to get to the point where any of this is relevant to US even if the worst that has been said turns out to be true.

All we know today is nothing less than we knew yesterday, which is that none of us are going to be buying a fermi-based anything from NV tomorrow.

Until I read a benchmark laden review from Anand I'm not really inclined to dwell on all the micro-milestone iterations that marketing tends to do with these things. I felt the same way about HD5870 too.

Tell me that whole eyefinity pre-view deal on 9/10 (or was it 9/12?) could not have waited for the actual release date to be done? And what value did it add to any of us to know about eyefinity a week or two before the card itself actually was available for purchase?

It is exciting and passes the time to reverse-engineer the milestones leading up to a product release, I partake in that myself and I hold it against no one if they do it either, just saying personally for me this round of GPU releases has left me uninspired. Not enough hype, too much hype, NV is too quiet, NV is trying to steal AMD's thunder, blah blah.

Unlike past releases, the build-up to this one has felt more like a waste of time than ever before. Could be I'm just getting older and I hold less value in this type of small stuff as it were.

 

[Tempered81]

What they showed was the CEO holding the GF100 in his hand (the board), and then before and after the keynote he held the Chip with an IHS on it for some press shots. During the opening keynote they showed on a projector a side-by-side comparison of a fluid benchmark being run on GT200 vs. GF100 architecture. The framerates for the demonstration (which looked like the valve particle benchmark and can be seen on youtube) were said to favor the GF100 5 to 1 or greater - due to performance increases in the GPGPU arena. They had some more demonstrations like a Laptop playing back HDTV and connecting to a remote rendering farm via cloud computing. These were the only hardware showings, the rest was speeches given by Jensen and speakers from Adobe and other VIP vendors. The only things they had directly related to GF100 was showing the board, showing the chip, and showing the Fluid bench.

 

[Kakkoii]

Originally posted by: Tempered81
What they showed was the CEO holding the GF100 in his hand (the board), and then before and after the keynote he held the Chip with an IHS on it for some press shots. During the opening keynote they showed on a projector a side-by-side comparison of a fluid benchmark being run on GT200 vs. GF100 architecture. The framerates for the demonstration (which looked like the valve particle benchmark and can be seen on youtube) were said to favor the GF100 5 to 1 or greater - due to performance increases in the GPGPU arena. They had some more demonstrations like a Laptop playing back HDTV and connecting to a remote rendering farm via cloud computing. These were the only hardware showings, the rest was speeches given by Jensen and speakers from Adobe and other VIP vendors. The only things they had directly related to GF100 was showing the board, showing the chip, and showing the Fluid bench.

They also showed an augmented reality demo, real time physics demo, and a ray-tracing demo. They rendered the ray-traced car in front of everyone.

http://forums.anandtech.com/me...=2339062&enterthread=y

 

[nitromullet]

Originally posted by: Idontcare

Originally posted by: nitromullet

Originally posted by: Idontcare
But if it makes you feel better to surmise that NV bandied about a "for display purposes only" tesla shell then be my guest, sure it is a fake. Now that we concluded and agreed to that, what does it change? About tesla? About GTX380? About anything?

Well, it might be relevant if it's an indicator of whether or not NV has working cards or not (Tesla or otherwise). Although, obviously even if NV showed off only a non-functional shell, it doesn't necessarily mean that they don't have cards working in their labs.

That would be my concern...we have to travel waaaay down the assumption train to get to the point where any of this is relevant to US even if the worst that has been said turns out to be true.

All we know today is nothing less than we knew yesterday, which is that none of us are going to be buying a fermi-based anything from NV tomorrow.

Until I read a benchmark laden review from Anand I'm not really inclined to dwell on all the micro-milestone iterations that marketing tends to do with these things. I felt the same way about HD5870 too.

Tell me that whole eyefinity pre-view deal on 9/10 (or was it 9/12?) could not have waited for the actual release date to be done? And what value did it add to any of us to know about eyefinity a week or two before the card itself actually was available for purchase?

It is exciting and passes the time to reverse-engineer the milestones leading up to a product release, I partake in that myself and I hold it against no one if they do it either, just saying personally for me this round of GPU releases has left me uninspired. Not enough hype, too much hype, NV is too quiet, NV is trying to steal AMD's thunder, blah blah.

Unlike past releases, the build-up to this one has felt more like a waste of time than ever before. Could be I'm just getting older and I hold less value in this type of small stuff as it were.

Agreed. Although, NV really sort of had to have something to say or show about its next generation this week. They couldn't really let ATI have a successful launch with availability, and not make a peep. I'm sure ATI will have something to say as well the week GTX380/Fermi/GF100 launches too.

Originally posted by: Kakkoii
They also showed an augmented reality demo, real time physics demo, and a ray-tracing demo. They rendered the ray-traced car in front of everyone.

http://forums.anandtech.com/me...=2339062&enterthread=y

Well, that pretty much makes this thread moot IMO. Regardless of whether or not the card Huang held in his hand was functional or not, NV appears to have Fermi running. So, who cares if the card works.

 

[Tempered81]

except all those GTC hardware demos you mention were run on GT200. 2:03 in the veyron demo, listen to jensen.

 

[Kakkoii]

Originally posted by: Tempered81

And why there is a 6-pin and 8-pin connector, but only 10 unmatching solder points on the PCB?
http://i34.tinypic.com/2d7yaew.jpg
http://i34.tinypic.com/2wmdfec.jpg

The card only uses an 8-Pin connector. They stated it consumes less energy than the GT200 series did. What looks like a 4 pin connector on the top isn't functional, you can see a traced outline of where the 4 solder points would have been on the back. Perhaps it's just junk from prototyping. But the required 8-pin is there and soldered. And that back plate helps keep micro-fractures from happening by reducing any bending of the card that might happen. At least that's what I'd assume. It probably has a cooling related purpose also.

 

[Kakkoii]

Originally posted by: Tempered81
except all those GTC hardware demos you mention were run on GT200.

Uhh.. No they weren't. They were run on Fermi. That was the whole point of showing them off.

 

[Tempered81]

goto 2:03 in the video you linked and turn your speakers up

 

[Keysplayr]

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

Jared, what is the point you'll be eventually getting to very soon I hope?
Fermi gets here when it gets here. If you're trying to pinpoint a launch or lack thereof, good luck.

 

[Kakkoii]

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

Ah I see for the ray-trace demo. It did seem the same speed as older demos.

 

[Tempered81]

I need to have reason to correct that person when they say "this is being run on GF100" when in fact it isn't? I can leave this topic alone from now on, but this is exactly the reason i started it - so you don't have guys like him saying,
"Well they showed GTX380 back in September and ran all those demos on it. It's now almost January 2010, and I still haven't seen a review on it, jeez what's wrong, I dont understand?"

Does that sound like something important for AT readers to know? Or should they just think they saw what we think it is. I was trying to provide some facts, so that everyone doesn't think a GF100 is already ray-tracing a Veyron.
I'll leave it alone though.

:disgust:

 

[nitromullet]

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

In this video: http://www.youtube.com/watch?v=W-9epnnDSUM

He states that it's "few day old silicon" running. Start at 3:10.

 

[Tempered81]

Originally posted by: nitromullet

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

In this video: http://www.youtube.com/watch?v=W-9epnnDSUM

He states that it's "few day old silicon" running. Start at 3:10.

Originally posted by: Tempered81
What they showed was the CEO holding the GF100 in his hand (the board), and then before and after the keynote he held the Chip with an IHS on it for some press shots. During the opening keynote they showed on a projector a side-by-side comparison of a fluid benchmark being run on GT200 vs. GF100 architecture. The framerates for the demonstration (which looked like the valve particle benchmark and can be seen on youtube) were said to favor the GF100 5 to 1 or greater - due to performance increases in the GPGPU arena. They had some more demonstrations like a Laptop playing back HDTV and connecting to a remote rendering farm via cloud computing. These were the only hardware showings, the rest was speeches given by Jensen and speakers from Adobe and other VIP vendors. The only things they had directly related to GF100 was showing the board, showing the chip, and showing the Fluid bench.

 

[novasatori]

Oh man, a fake card.

All the specs are null and void now, this card will never come out - I had high hopes for it too....

 

[nitromullet]

Originally posted by: Tempered81

Originally posted by: nitromullet

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

In this video: http://www.youtube.com/watch?v=W-9epnnDSUM

He states that it's "few day old silicon" running. Start at 3:10.

Originally posted by: Tempered81
What they showed was the CEO holding the GF100 in his hand (the board), and then before and after the keynote he held the Chip with an IHS on it for some press shots. During the opening keynote they showed on a projector a side-by-side comparison of a fluid benchmark being run on GT200 vs. GF100 architecture. The framerates for the demonstration (which looked like the valve particle benchmark and can be seen on youtube) were said to favor the GF100 5 to 1 or greater - due to performance increases in the GPGPU arena. They had some more demonstrations like a Laptop playing back HDTV and connecting to a remote rendering farm via cloud computing. These were the only hardware showings, the rest was speeches given by Jensen and speakers from Adobe and other VIP vendors. The only things they had directly related to GF100 was showing the board, showing the chip, and showing the Fluid bench.

So, you already know they have Fermi running... What's the problem?

 

[Kakkoii]

Originally posted by: Tempered81

Originally posted by: nitromullet

Originally posted by: Tempered81
goto 2:03 in the video you linked and turn your speakers up

In this video: http://www.youtube.com/watch?v=W-9epnnDSUM

He states that it's "few day old silicon" running. Start at 3:10.

Originally posted by: Tempered81
What they showed was the CEO holding the GF100 in his hand (the board), and then before and after the keynote he held the Chip with an IHS on it for some press shots. During the opening keynote they showed on a projector a side-by-side comparison of a fluid benchmark being run on GT200 vs. GF100 architecture. The framerates for the demonstration (which looked like the valve particle benchmark and can be seen on youtube) were said to favor the GF100 5 to 1 or greater - due to performance increases in the GPGPU arena. They had some more demonstrations like a Laptop playing back HDTV and connecting to a remote rendering farm via cloud computing. These were the only hardware showings, the rest was speeches given by Jensen and speakers from Adobe and other VIP vendors. The only things they had directly related to GF100 was showing the board, showing the chip, and showing the Fluid bench.

The fact of the matter is, they have working silicon to run those demo's and then show the clips to the people. Doesn't matter if they had a working card there or not. They have working silicon, that's what matters most.

 

[Tempered81]

nvm.

 

[Idontcare]

Originally posted by: Tempered81
I need to have reason to correct that person when they say "this is being run on GF100" when in fact it isn't? I can leave this topic alone from now on, but this is exactly the reason i started it - so you don't have guys like him saying,
"Well they showed GTX380 back in September and ran all those demos on it. It's now almost January 2010, and I still haven't seen a review on it, jeez what's wrong, I dont understand?"

Does that sound like something important for AT readers to know? Or should they just think they saw what we think it is. I was trying to provide some facts, so that everyone doesn't think a GF100 is already ray-tracing a Veyron.
I'll leave it alone though.

:disgust:

Phrases like "much ado about nothing" and "making mountains out of molehills" exist for reason.

Your points are valid, not that there was any question in that so really there is no need for it to be reaffirmed, and with merit. But it seems like this is just taking it to an extreme that in an of itself is making people ask "so?"...which is the part where we are supposed to naturally self-assess and honestly ask ourselves "am I making a mountain out of a molehill here?".

It is your thread, you put in the effort, if you feel it IS a mountain then godspeed to ya man.

But don't begrudge us of the option of holding to the opinion this is a molehill at best, for now, we reserve the right to change our opinion of course as time passes and more facts develop.

 

[thilanliyan]

nm

 

[ronnn]

Has Jenson got Alzheimers? Has been quite a string of bizzare behavior. Not sure how he thought waving around a fake would not be noticed.

Maybe nvidia can invade the compute market and displace intel cpu's, but risky and doesn't do much for graphics.

 

[lopri]

nm, too.