Shotglass01
[H]ard|Gawd
- Joined
- Aug 26, 2005
- Messages
- 1,992
Yes, it's Fudzilla. So take with a healthy dose of salt. But here it is:
http://www.fudzilla.com/content/view/16185/1/
http://www.fudzilla.com/content/view/16185/1/
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Fudzilla Says GF100 in Early Dec.
- Thread starter Shotglass01
- Start date
Doesn't sound completely unrealistic, but we'll have to see 
I'll keep an eye on any other rumours on this I may encounter on more reputable sites.
And to think I just bought a GTX260... ah, buyer's remorse is always around the corner
I'll keep an eye on any other rumours on this I may encounter on more reputable sites. And to think I just bought a GTX260... ah, buyer's remorse is always around the corner
jeremyshaw
[H]F Junkie
- Joined
- Aug 26, 2009
- Messages
- 12,511
It's always around the corner. Just wait a few months, and something new will start tugging.
Also, a salt overdose was once an accepted way of suicide in Japan. So.... no 'healthy' dose?
Also, a salt overdose was once an accepted way of suicide in Japan. So.... no 'healthy' dose?
Sodium's an essential mineral, so there could be a healthy dose depending on the rest of one's diet.
l88bastard
2[H]4U
- Joined
- Oct 25, 2009
- Messages
- 3,734
It says they are taking Pre-orders. Where do you pre-order it?
Sounds like the A typical ATI launch so all good there.
When was the last time ATI did a paper launch?
There was availability of the 5800, 4800, and 3800 series from my own memory.
Or maybe we are using different definitions of paper launch?
http://en.wikipedia.org/wiki/Paper_launch
I believe Nvidia is taking pre orders from its partners, not end users.
Brackle
Old Timer
- Joined
- Jun 19, 2003
- Messages
- 8,585
As Prime1 said on techreport
"Finally a real DX11 card"
Just in before he says it
"Finally a real DX11 card"
Just in before he says it
if they can launch this is probably it. I am sure they have working silicon by now but TSMC just admitted that their yields just went from 60 to 40 %. and that was for the ATI chips. and they can't produce enough of them. Still even if they send them out to reviewers the results should be interesting.
D
Deleted member 174368
Guest
So are the 5800's "fake" DX11 cards?
What the hell is that statement supposed to mean
edit: I know PRIME1 is a crazy nvidia fanboy, but that's a bit too much even for him.
Brackle
Old Timer
- Joined
- Jun 19, 2003
- Messages
- 8,585
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,123
Its probably EVGA, BFG etc. If there were any consumers that could do preordering, the price point would be interesting news for the new cards. Perhaps the whole thing is FUD on a slow newsday...
Would be nice if they did come out with a release date for general availability soon though.
Hehehe... Prime1 is supposedly Wreakage@anandtech. I've had lots of laughs reading his posts. Especially the one where he found a patent of a tessellation method belonging to Nvidia from 2003. He got a lot of pepper for it, but I think it was cute he tried to find some tessellation info about Nvidia with all the talk about DX11 tessellation these days.
He doesn't have to worry, Nvidia is supposed to support DX11 fully with Fermi and thus tessellation as well.
I wouldn't be suprised if he one day would be the founder of "The church of Nvidia" though. Some of his posts are almost religious already. LOL!
Lorien
Supreme [H]ardness
- Joined
- Aug 19, 2004
- Messages
- 5,197
Tessellation without dedicated hardware will run slower because it is all being handled in software. It will reduce framerate just like physx does when running on the gpu. If you doubt this, ask yourself why did evga just release a gtx275/gts250 hybrid with one gpu locked out and dedicated solely for physx. Expect a sky high smoke screen from Nvidia knocking down tessellation and hyping physx.
Brackle
Old Timer
- Joined
- Jun 19, 2003
- Messages
- 8,585
Its probably EVGA, BFG etc. If there were any consumers that could do preordering, the price point would be interesting news for the new cards. Perhaps the whole thing is FUD on a slow newsday...
Would be nice if they did come out with a release date for general availability soon though.
Hehehe... Prime1 is supposedly Wreakage@anandtech. I've had lots of laughs reading his posts. Especially the one where he found a patent of a tessellation method belonging to Nvidia from 2003. He got a lot of pepper for it, but I think it was cute he tried to find some tessellation info about Nvidia with all the talk about DX11 tessellation these days.
He doesn't have to worry, Nvidia is supposed to support DX11 fully with Fermi and thus tessellation as well.
I wouldn't be suprised if he one day would be the founder of "The church of Nvidia" though. Some of his posts are almost religious already. LOL!
Yea it was a pretty good laugh thats for sure.
Lets hope Fermi is all its suppose to be, then prices will go down
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,123
Yea it was a pretty good laugh thats for sure.
Lets hope Fermi is all its suppose to be, then prices will go down
I have no doubt that Nvidia will bring something worthwhile to the market. They wouldn't go this route with Fermi, unless they thought it would be a market for it and money could be made.
At least I expect it to be able to compete in games, which will bring prices down on GPU's as you say. Its not good for us consumers that only ATI have a DX11 card out now.
Fermi is also supposed to have full support for DX11, which should include hardware support for tessellation. Anything else would be strange.
As for physx, it will probably get a harder battle with Bullet and Havok physics next year and also against DX11. Some have strangely enough been so impressed with waving banners, dust, water and particles in the air and that you got smoke with Batman only with PhysX... Well, Dirt2 has it all in DX11. I've seen those things in other games before, but some might find it impressing then in Dirt 2 considering its done without physx:
Dirt 2 DX11 tech demo:
http://www.youtube.com/watch?v=D9p3PYOX1Vc
Dirt 2 DX11 gameplay demo:
http://www.youtube.com/watch?v=-re-oaYaCdg
Hopefully Nvidia will make physx on opencl or compute shaders to give it something worthwhile to games that cannot be done without it in DX11. Then I might be easier sold on physx.
Lorien
Supreme [H]ardness
- Joined
- Aug 19, 2004
- Messages
- 5,197
Wrong, read the white paper. There is no dedicated tessellation hardware on Fermi.
Hopefully Nvidia will make physx on opencl or compute shaders to give it something worthwhile to games that cannot be done without it in DX11. Then I might be easier sold on physx.
They won't. Did you forget that they purposely rendered their current customers' hardware useless with their drivers if they happen to want to use the competition as the main renderer. They defended this with the usual PR spin BS to "protect" physx.
Even so, this anti-consumer stunt won't make much of a difference in the long run. The industry will either migrate to open standards or someone will write a physx wrapper to run in OpenCL.
There's been dedicated tesselation hardware on GPUs for years already, just that only OpenGL has been using it so far because MSFT deemed it 'unsuitable' for use with D3D.
*shrugs*
I have no doubt that Nvidia will bring something worthwhile to the market. They wouldn't go this route with Fermi, unless they thought it would be a market for it and money could be made.
At least I expect it to be able to compete in games, which will bring prices down on GPU's as you say. Its not good for us consumers that only ATI have a DX11 card out now.
I am sure they can, what worries me is that they are worried more about the HPC market then the gamers. The way they wrote the white papers it almost like "yeah, it can render graphics too" (maybe I am paranoid) Hopefully they have their foots in with GF as well, TSMC is starting to look like its just isn't going to be able to deliver here
5800 was about as paper as it comes. What 2 months later and still can't find one of these cards on newegg to save your life. If you scratch around on the net you can find them here and there at mom and pops but otherwise: PAPER LAUNCH. Anyone that refutes that is a phan.
Heh, I guess Astrodave has missed the previous ATi and nVidia launches where nobody even saw the hardware in some obscure Asian store let alone at a major retailer until months after the 'launch'. Being able to get the card with difficulty most likely means it's still not being produced in sufficient quantities and/or demand is really high. Kind of like the Wii launch
except a lot of people have them here. a paper launch is when you announce a product that you don't have to sell. they did and have, in fact selling out as soon as they get stocked. with the TSMC issues I would expect much the same for Nvidia. but they did and do have stock if not enough to satisfy everyone. Still they should have forseen this (I know they can't control TSMC issues) and gotten a larger stockpile before hand rather then expecting the manufactures to keep up
D
Deleted member 174368
Guest
So how come I, and many other forum members, own one?
You're confusing "paper launch" with "high demand". The cards sell out within 5 minutes of being in stock. Not to mention Dell bought a huge amount of them to use in their new Alienware systems.
By your logic, the Wii was also a paper launch since it wasn't easy to find one for over a year after it was released.
you can call it whatever you want, you can blame TSMC for not getting their shit together, they said their 40nm production was fixed and now we hear that it is gotten worse again, down to 40% yields so if it was ati or nvidia they would both be going through the same thing regardless of how hard they try, can't wait until 5 years from now where amd produces their own gpu's and their own cpu's in their own fab.
I mean I always saw them available when I got mine, there was one brand available or the other almost everyday, but now it seems to have gotten a lot worse, TSMC seems to explain why.
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,123
That might be so, but it might perform well even so without a fixed tessellation hardware on GPU.
Its consumer hostile the way Nvidia have implemented physx now with blocking even their own hardware, this we agree upon. But, physx is not the only game in town with hardware accelerated physics anymore (after port of Bullet to both DX11 and Opencl). Nvidia might be forced to change their strategy for the survival of physx.
Fermi was directed towards the HPC market in beginning of october. I think they needed to start early there, gaining some credibility there. They even advertised one company signing on for a supercomputer with Fermi's, though even the Nvidia CEO had to use a fake Fermi card for display there. Usually companies wait to see how they work, before making such large purchase, so either Nvidia was dealing with a fool or someone got a golden deal. But, the HPC market and the gaming market are very different segments. Therefore, I think that Nvidia had sole focus on HPC that round and will focus more for gamers when the consumer version is closer to launch.
Because it would make perfect sense to hand an expensive prototype card over to the CEO and have lots of journalists fondle it with their grubby hands, ensuring it will most decidedly never work again afterwards. The actual demos during the show were run on Fermi hardware.
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,123
Nobody was fondleling with the fake card either. (Its funny though that the CEO presented the card as the real one. That must have been embarrasing afterwards once it was discovered).
Thats beside the point though. My point was that they advertised that a company had signed up for next years supercomputer already then and thats without having even some working cards for the company to test first. For a company to make such investment for next year without any tests for stability, would mean that the ones responsible for buying the supercomputer setup (which is extremely expensive usually) an idiot, or they got a sweet deal to be a posterboy for HPC. Since it takes time usually for companies to make such investments before its tried and tested much, I would guess that Nvidia wanted to make focus on the HPC capabilities early so to give those companies time.
This is why I assume that their HPC display wasn't because they are abandoning gaming market, but that Nvidia needed to announce their new HPC capable Fermi earlier, so companies will get time to evaluate it on an earlier stage. The Fermi announced with ECC memory and such is not supposed to be the consumer version, which I belive will be announced with its specs and prices closer to launch.
Anyone that calls this a paper launch is full of sour grapes and butt hurt from not getting simplistic AMD drivers working.
needmorecarnitine
Supreme [H]ardness
- Joined
- Feb 10, 2004
- Messages
- 4,609
Nobody was fondleling with the fake card either. (Its funny though that the CEO presented the card as the real one. That must have been embarrasing afterwards once it was discovered).
Thats beside the point though. My point was that they advertised that a company had signed up for next years supercomputer already then and thats without having even some working cards for the company to test first. For a company to make such investment for next year without any tests for stability, would mean that the ones responsible for buying the supercomputer setup (which is extremely expensive usually) an idiot, or they got a sweet deal to be a posterboy for HPC. Since it takes time usually for companies to make such investments before its tried and tested much, I would guess that Nvidia wanted to make focus on the HPC capabilities early so to give those companies time.
This is why I assume that their HPC display wasn't because they are abandoning gaming market, but that Nvidia needed to announce their new HPC capable Fermi earlier, so companies will get time to evaluate it on an earlier stage. The Fermi announced with ECC memory and such is not supposed to be the consumer version, which I belive will be announced with its specs and prices closer to launch.
You are making a lot of assumptions. How can you be so sure nVidia by now doesn't have piles of prototypes and haven't tested the heck out of it already? They first start by doing hardware simulations to get the worst bugs, once the first ASICs are put together there shouldn't be any critical bugs left in it. Since the demos at the show were running on prototypes (i.e. real hardware) it's reasonable to assume that at this point nVidia does have more or less finalized hardware, just needs to finish the flashy bits like designing the final PCB layout and a cool looking cooler/shroud. Since the media and the average person will start yelling that nVidia is a liar and the company doomed if nVidia doesn't show them something they expect to see (a final-looking PCB and shroud, not a debug-enabled prototype with wires sticking from underneath a functional-but-ugly cooler), they had little choice but to do what they did.
The people who actually give them money (card manufacturers) will have access to those prototypes and will actually know WTF they're talking about. Since nVidia doesn't produce cards themselves I'm still not sure WTH people are getting so bloody excited about. For crying out loud.
Hence my sarcastic post above. Which btw I knew would get the fanboy's panties in a bunch. I had my popcorn out. LOL
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,123
What they have by now is irrelevant. What they had when a company signed up for a supercomputer with Fermi is what I am talking about (for crying out loud).
Of course I assume and so do you. Thats why I write "This is why I assume that their HPC display wasn't because they are abandoning gaming market". I don't know what the heck you are discussing. Do you disagree with my assumptions or are you just discussing wheter or not Nvidia is doomed because they presented a fake card and told it was real at the conference?
If you disagree on the first part, please argue why you think they are abandoning the gaming market.
For the second part, I don't think they are doomed because they lied. I do however think that the company that bought a supercomputer based on, as you say: "a debug-enabled prototype with wires sticking from underneath a functional-but-ugly cooler", must have been an idiot or have gotten a golden deal to become a posterboy for Fermi HPC.
As for your "they had little choice but to do what they did.". Yes, they did have the choice of not lying into people's faces.
Here, the CEO of Nvidia is showing off the card. Do you actually support his actions of taking out the fake card from the cloth and presenting it as real? Lying to the crowd? It doesn't look to me as he had a gun to his head and no choice.
http://www.youtube.com/watch?v=bJJyG67by0U
The whitepaper also doesn't mention rasterization, triangle setup or texturing. So drawing conclusions about tessellation from the white paper isn't really a worthwhile exercise.