New Nvidia Rumors?

[Seero]

NVIDIA's CEO stated at CES that the cards are already in production.

What I said doesn't contradict with that. The cards are there, but the clock speed is not determined yet.

 

[Meghan54]

Fermi is a new architecture. The 5xxx series is not and that's why the 5870 gets beat by an old GTX295. It's why the 5xxx series is not significantly faster than the GT2xx series.

You're right, Wreckage, NVIDIA still has the faster single-card solution, although the $120 price premium is well in excess of the <10% performance premium. (Comparing the current price for 5870 at $399 vs. a completely unavailable GTX295 at $520.)

 

[dguy6789]

Fermi is a new architecture. The 5xxx series is not and that's why the 5870 gets beat by an old GTX295. It's why the 5xxx series is not significantly faster than the GT2xx series. Probably why they lost marketshare last quarter.

Either way I'm just going by the information that came out in and around CES. Not my own personal feeling on the issue.

So you have no basis what so ever for your claims? Figured as much but I asked just to make sure.

 

[v8envy]

It might be possible to do a "press edition" dual card no matter how power hungry Fermi is. Odds are very high NV will be able to find some cores capable of running wanted clocks at half the current (see: my i7 920, happily running 2.8Ghz on .89v) to make enough for the press.

The real question will be: can you buy them retail?

 

[imported_NoGodForMe]

The other thing to talk about is PCIe 3.0 due out this year. Read the Wiki and it puts things into perspective.
http://en.wikipedia.org/wiki/PCI_Express
The spec is due out in Q2 2010. It would be a bummer to build a new machine with Fermi (or whatever), then have PCIe 3.0 come out with new cards that use it. I already ran into that with my previous build, put in a water cooled AGP card, and then PCI came out. AGP was practically dropped in terms of faster cards. That's not going to happen to me again. I'll wait for PCIe 3.0 even if it's later this year.

PCIe 3.0 is a much bigger event than USB 3.0, or Sata 6Gb.

 

[scooterlibby]

The other thing to talk about is PCIe 3.0 due out this year. Read the Wiki and it puts things into perspective.
http://en.wikipedia.org/wiki/PCI_Express
The spec is due out in Q2 2010. It would be a bummer to build a new machine with Fermi (or whatever), then have PCIe 3.0 come out with new cards that use it. I already ran into that with my previous build, put in a water cooled AGP card, and then PCI came out. AGP was practically dropped in terms of faster cards. That's not going to happen to me again. I'll wait for PCIe 3.0 even if it's later this year.

PCIe 3.0 is a much bigger event than USB 3.0, or Sata 6Gb.

I am not trying to be snarky, but why would that matter initially? Does even the mighty 5970 saturate PIC-E 2.0x16 yet?

 

[Wreckage]

So you have no basis what so ever for your claims? Figured as much but I asked just to make sure.

And your information comes from where?

You can read the CES thread and the benchmarks that came out. Search this very forum.

Take your FUD elsewhere.

 

[Seero]

The other thing to talk about is PCIe 3.0 due out this year. Read the Wiki and it puts things into perspective.
http://en.wikipedia.org/wiki/PCI_Express
The spec is due out in Q2 2010. It would be a bummer to build a new machine with Fermi (or whatever), then have PCIe 3.0 come out with new cards that use it. I already ran into that with my previous build, put in a water cooled AGP card, and then PCI came out. AGP was practically dropped in terms of faster cards. That's not going to happen to me again. I'll wait for PCIe 3.0 even if it's later this year.

PCIe 3.0 is a much bigger event than USB 3.0, or Sata 6Gb.

The specification will probably be finalized at Q2 2010. Actual products are probably going to be out next year, 2011. Keep waiting.

 

[Daedalus685]

The other thing to talk about is PCIe 3.0 due out this year. Read the Wiki and it puts things into perspective.
http://en.wikipedia.org/wiki/PCI_Express
The spec is due out in Q2 2010. It would be a bummer to build a new machine with Fermi (or whatever), then have PCIe 3.0 come out with new cards that use it. I already ran into that with my previous build, put in a water cooled AGP card, and then PCI came out. AGP was practically dropped in terms of faster cards. That's not going to happen to me again. I'll wait for PCIe 3.0 even if it's later this year.

PCIe 3.0 is a much bigger event than USB 3.0, or Sata 6Gb.

That is when they will finalize the spec itself. We won't see anything using it until next gen at the earliest. Unless of course it has higher power accepted levels and ATI/Nvidia release 5970s/395s under that spec just for power requirements alone.

For instance, USB 3.0 was finalized in 2008 and we are just seeing devices for it now.

 

[Munky]

The other thing to talk about is PCIe 3.0 due out this year. Read the Wiki and it puts things into perspective.
http://en.wikipedia.org/wiki/PCI_Express
The spec is due out in Q2 2010. It would be a bummer to build a new machine with Fermi (or whatever), then have PCIe 3.0 come out with new cards that use it. I already ran into that with my previous build, put in a water cooled AGP card, and then PCI came out. AGP was practically dropped in terms of faster cards. That's not going to happen to me again. I'll wait for PCIe 3.0 even if it's later this year.

PCIe 3.0 is a much bigger event than USB 3.0, or Sata 6Gb.

PCI-e is backwards compatible with previous versions, this is not like AGP vs PCI-e. I ran a 2.0 card in a 1.x slot without problems, don't see why you can't run a future 3.0 card in a 2.0 slot. And you could most definitely run a 2.0 card in a 3.0 slot.

 

[dguy6789]

And your information comes from where?

You can read the CES thread and the benchmarks that came out. Search this very forum.

Take your FUD elsewhere.

Woah woah woah. You made some incredibly ridiculous claims and I asked you to back them up. I don't know how you can possibly think I need to prove anything. This discussion is over.

 

[Daedalus685]

And your information comes from where?

You can read the CES thread and the benchmarks that came out. Search this very forum.

Take your FUD elsewhere.

Rational pessimism is not FUD....

 

[Janooo]

Fermi is a new architecture. The 5xxx series is not and that's why the 5870 gets beat by an old GTX295. It's why the 5xxx series is not significantly faster than the GT2xx series. Probably why they lost marketshare last quarter.

Either way I'm just going by the information that came out in and around CES. Not my own personal feeling on the issue.


NVIDIA's CEO stated at CES that the cards are already in production.

Who lost the market share? AMD? Can you prove it?

 

[SHAQ]

I am not trying to be snarky, but why would that matter initially? Does even the mighty 5970 saturate PIC-E 2.0x16 yet?

It probably does, but we won't know until they run tests with it on 3.0.

 

[happy medium]

I am not trying to be snarky, but why would that matter initially? Does even the mighty 5970 saturate PIC-E 2.0x16 yet?

Well the 5870 doesn't saturate the pci-e 16 1.0 bus ,and the 5970 is slower then 2 5870's, so the answer is NO the 5970 does not saturate the pci-e 2.0 x16 bus.

 

[imported_NoGodForMe]

For instance, USB 3.0 was finalized in 2008 and we are just seeing devices for it now.

So I guess I'm over reacting, that's what I was trying to find out. And I agree that all cards will be compatible (3.0 with 2.0 and so on). It was just such a bummer to build that AGP system and then have no upgrade after a couple years. Yes, I could have bought a new MB, CPU, memory, GPU combination, but that usually means a re-format, and that's practically the entire computer at that point. That's why you see so many people who waited 5 years, who had a 939 system, just like me.

 

[ShadowOfMyself]

You're right, Wreckage, NVIDIA still has the faster single-card solution, although the $120 price premium is well in excess of the <10&#37; performance premium. (Comparing the current price for 5870 at $399 vs. a completely unavailable GTX295 at $520.)

Hah? When did the 5970 cease to exist?

And Lol @ 5870 being slower than the "old 295", the GTX285 was much slower than 4870X2, but I dont remember you (Wreckage) bringing up that point, I wonder why

It only matters if its a dual gpu when its the opposite company doing it

 

[waffleironhead]

Who lost the market share? AMD? Can you prove it?

"AMD gained in the notebook integrated segment, but lost some market share in discrete in both the desktop and notebook segments due to constraints in 40nm supply. Nvidia picked up a little share overall."

http://jonpeddie.com/press-releases...in-pc-graphics-quarter-to-quarter-also-beats/

amd dropped from 20.1% to 19.9%, but the author claims it was due to 40nm supply constraints, NOT performance, while nV went from 25.3 to 24.3 so they gained slightly in AIB and lost overall as well.

So who did they both lose to? Intel.

 

[Wreckage]

And Lol @ 5870 being slower than the "old 295", the GTX285 was much slower than 4870X2, but I dont remember you (Wreckage) bringing up that point, I wonder why

It only matters if its a dual gpu when its the opposite company doing it

Your attempt to discredit me has failed. A better comparison would be the GTX285 vs the 3870x2.

Just as I am comparing the 5870 vs the GTX295.

Pretty simple really.

 

[kalrith]

I would definitely wait until after fermi is release. We should see a few price dips, plus, if it is a winner you'll always have the option to go with it over some other video card.

This is what I'm doing. Regardless of whether I go with nvidia or ATI I'll get more for my money if I wait until after Fermi's released.

 

[akugami]

And your information comes from where?

You can read the CES thread and the benchmarks that came out. Search this very forum.

Take your FUD elsewhere.

The benchmarks are more in the way of previews. Press release benchmarks, regardless of company, tend to be deceptive. In other words, you still have zero basis for your claims of how fast Fermi is. From your other threads, you're a stickler for stating fact and opinion. Even though the writer knows when he's stating an opinion you have to point it out anyways. Probably in case some poor reader can't tell the difference between a fact and an opinion.

Since you're such a stickler for facts and opinions, the question is where do you get some of your info, such as a dual GPU Fermi being out in April? Which benchmarks are you using when making the speed/power claims concerning Fermi vs Radeon 5xxx?

Your claims about the Radeon 5 series vs the Geforce 200 series are also highly subjective and more along the lines of opinion. It's also inconsistent. You're going to compare a 5870 which is a single GPU video card, to a Geforce 295 which is a dual GPU card? At least compare it to a 5970.

 

[evolucion8]

Your attempt to discredit me has failed. A better comparison would be the GTX285 vs the 3870x2.

Just as I am comparing the 5870 vs the GTX295.

Pretty simple really.

No, you had done an excellent job discrediting yourself spreading lies and biased opinions like the following one;

Fermi is a new architecture. The 5xxx series is not and that's why the 5870 gets beat by an old GTX295. It's why the 5xxx series is not significantly faster than the GT2xx series. Probably why they lost marketshare last quarter.

Either way I'm just going by the information that came out in and around CES. Not my own personal feeling on the issue.


NVIDIA's CEO stated at CES that the cards are already in production.

The HD 5870 is only 10% slower than the GTX 295 in the worst case scenario, usually both are very close, are you comparing two GPU's against one? Please, even the HD 4870X2 comes so close to the GTX 295 quite often that it made it irrelevant for a while.

So what you posted there was a biased personal feeling of favoritism to nVidia which all people here on these forums know and which made you unpopular. What market share has to do with the topic? nVidia lost market share this last quarter too! Using your mentality I would state that nVidia lost market share because of its constant rehashes thinking that their customers are a bunch of idiots? Definitively is not a smart post by any means.

 

[Wreckage]

The HD 5870 is only 10% slower than the GTX 295 in the worst case scenario,

http://www.techreport.com/articles.x/17618/10

It took me barely 2 seconds to prove you wrong. So now who is lying?

D:

 

[SlowSpyder]

The 58xx is as new as Fermi is going to be. AMD seems to tweak/add to their architectures as they go, supporting DX10.1, building with 40nm first, DDR5, using a tesselator first, etc. Nvidia HAS to more heavily work Fermi from where the GT200 cards are to get Fermi to market. AMD already had the groundwork layed down.

And you can try to deflect 58xx to GT2xx comparrisons all you want, but when they occupy the same price point then to most people that's where they compete.

 

[evolucion8]

http://www.techreport.com/articles.x/17618/10

It took me barely 2 seconds to prove you wrong. So now who is lying?

D:

Yeah, a cheap labor. Wolfenstein is a CPU limited game that doesn't challenge enough the GPU. Always posting the same thing from Techreport, are u in love with that page? I wonder why.

From your own link;

http://www.techreport.com/articles.x/17618/16

The Radeon HD 5870 is the fastest GPU on the planet, with the best visual output, and the most compelling set of features. Yet it's still a mid-sized chip by GPU standards.

Look at the difference in scores between Anandtech and Techreport regarding Far Cry 2

http://www.anandtech.com/video/showdoc.aspx?i=3643&p=18

http://www.techreport.com/articles.x/17618/9

Anandtech review shows more realistic results with a faster setup, Techreport is mediocre at best, so your posts.