Nvidia confirms GT-200 with 1 billion transistors

[RealTelstar]

I'm actually expecting a 512bit bus.

Me too. Hoping for 1gig ram as well.

[RealTelstar]

Anyway, I'm hearing a late may-early june release, just thought I'd add that.

That sounds yummy

[RealTelstar]

The interesting part will be the high resolution & details with AA/AF, where the card will have a greater advantage over resolutions lower than 1680x1050 ( when compared to the GX2 )

Exactly my point. New games which uses a lot of textures are crippled by only 512mb per core that the g92 has.

This can be surpassed already by the g98 or whatever is the Q2 refresh that nvidia is holding for us.

(i corrected your resolution typo in the quoted text)

[BenchZowner]

LoL and this is where I call BS on you.

Good to hear that

You may know what Nvidia is releasing but you have no idea what AMD will compete with.

How come ? Are you inside my brain now ?
I know things, but not enough to have a crystal clear image, or "allowance" to mention more things that I did.

You make it sound like AMD NEEDS more ROPs.

If you want to game with 4xAA + you NEED them.

HD4870x2 may or may not be able to compete with GT200 but it should be a very interesting product and I have a feeling it might surprise some people.

Everything's interesting. Even a new revolutionary light bulb is interesting.
Thing is, with what I've seen we're about to face another Anti-Aliasing & Anisotropic Filtering limited card, yet with worse IQ than G8x/G9x/GT200.

Biggest issue in this situation is AMD's economy.
nVIDIA is playing "wait & strike" with AMD. If the HD4870X2 is good enough to beat let's say the 9800GX2, they can release a new card derived from GT200.
Even a new G92b dual card will more than likely do the trick.

Call me whatever you like, fanboy ? ? BS ?
I want ATi to be really competitive again ( if not surpass ) with nVIDIA, but the lack of a real high end part ( yet again ) and the architecture enhancements they seem to carry on the new HD4xxxx series... doesn't look good and surely not revolutionary.
We're bashing nVIDIA for their transitions from 65nm to 55nm with G92b.
On the other side ATi is doing the same ( HD2900 -> HD3xxx -> HD4xxx ) while they "can't afford it" because they're already lagging behind, they're not having a performance and/or IQ advantage like nVIDIA to have the ability to release new stuff that differ just a tiny bit from their "predecessors".

[flopper]

Everything's interesting. Even a new revolutionary light bulb is interesting.
Thing is, with what I've seen we're about to face another Anti-Aliasing & Anisotropic Filtering limited card, yet with worse IQ than G8x/G9x/GT200.

Biggest issue in this situation is AMD's economy.
nVIDIA is playing "wait & strike" with AMD. If the HD4870X2 is good enough to beat let's say the 9800GX2, they can release a new card derived from GT200.
Even a new G92b dual card will more than likely do the trick.

amd´s economy affects them, but if the cards are beter than current and as priceworthy, they are good cards and enough for the year to come.
its not we need that much horsepower

[YukonTrooper]

amd´s economy affects them, but if the cards are beter than current and as priceworthy, they are good cards and enough for the year to come.
its not we need that much horsepower

You forget how the market works.

[Nuker_]

Don't forget it's only one billion trans. That is not upped much from the previous generation, so they cant' cramp everything you have mentioned with such a tight transistor budget.

[Cooper]

It's really funny when ppl saying they know something about either of the card

[Decami]

Why is nVidia rushing all these new products when most games in 2008 will be console ports, and current hardware should run them just fine milking people for money?

i dont know what console ports you have been playing, cause most run like crap on current computer hardware, freaking idiots never take the time to optimize, cause they have the other marketing idiots rushing with dollar signs in their eyes.

aside of coarse from the hardware most of us here at XS use, which if we had to explain what we use and how we have tweaked it to the normal person, it would probably make their head explode.

Let me also point out that while I usually lean towards Nvidia, ATI releases perfectly good cards, many people around here use ATI cards obviously. ATI's problem is not, not being able to release good products, their problem is they are not able to release "omg what the heck is that in my ac duct, hope it doesnt eat me!!" cards, as Nvidia does. ATI has problems (now) competing with Nvidias Flagships.

in lamens terms. I look at (and use) and Nvidia card and think, omg, holy crap, i need a cage for this thing!!!
then i look at (and use) ATI cards and I think, good price for performance, this.

[grimREEFER]

so when are the leaked benches from chinese websites gonna be out?
early june?

[adamsleath]

3$ says b4 end of May...for rv770.
typical blockbuster weekly dvd rental for that.

[nemsis]

so when are the leaked benches from chinese websites gonna be out?
early june?

thats a very good Question

[perkam]

3$ says b4 end of May...for rv770.
typical blockbuster weekly dvd rental for that.

Funny, the your post doesn't concern itself with the topic of the thread...



Perkam

[shiznit93]

BenchZowner you have gotten my hopes up, if the card doesn't come soon or isn't as powerful as you said my little framerate czar will be gunning for you

[LordEC911]

I am amazed you are still responding... Can't we just agree to disagree?

How come ? Are you inside my brain now ?
I know things, but not enough to have a crystal clear image, or "allowance" to mention more things that I did.

No, I am not in your head... I am just reading what you post.
I read between the lines of your posts and pick up some interesting tidbits.

If you want to game with 4xAA + you NEED them.

Really?
This is why it makes me so hard to believe anything you say when you say hilarious things like this.
AA is done on the shaders. 480SPs clocked at +900mhz will be fine.
16ROPs clocked at +900mhz w/ tweaks will also be fine.
Please remember what G92 is doing with "only" 16ROPs...

Thing is, with what I've seen we're about to face another Anti-Aliasing & Anisotropic Filtering limited card, yet with worse IQ than G8x/G9x/GT200.

How do you know this?
R600 IQ is on par with G80/G92, R600 has better AA and worse AF.
I would consider that a wash.
I don't see RV770 being AA limited, especially when it, probably, shouldn't be used on a monitor higher than 1680x1050. I also don't see it being obviously held back in the AF department, at least doubling the TMU/TFUs and tweaking some other things should alleviate the lack of texturing power that has been haunting the architecture.

Biggest issue in this situation is AMD's economy.
nVIDIA is playing "wait & strike" with AMD. If the HD4870X2 is good enough to beat let's say the 9800GX2, they can release a new card derived from GT200.
Even a new G92b dual card will more than likely do the trick.

Call me whatever you like, fanboy ? ? BS ?
I want ATi to be really competitive again ( if not surpass ) with nVIDIA, but the lack of a real high end part ( yet again ) and the architecture enhancements they seem to carry on the new HD4xxxx series... doesn't look good and surely not revolutionary.
We're bashing nVIDIA for their transitions from 65nm to 55nm with G92b.
On the other side ATi is doing the same ( HD2900 -> HD3xxx -> HD4xxx ) while they "can't afford it" because they're already lagging behind, they're not having a performance and/or IQ advantage like nVIDIA to have the ability to release new stuff that differ just a tiny bit from their "predecessors".

Here is another spot where I have to disagree with you.
AMD has some OBVIOUS advantages go into this next round and while they may or may not take the performance crown, they will certainly be MORE than competitive with whatever Nvidia can throw at them.

[Barys]

Heh some guy on Beyond3d Forum have said that GT200 is THREE times faster than GF8800GTX in 3D Mark Vantage Could it be true? I`d expect something about 2 times faster maybe 2,5 in some situations but 3 times? It`s impossible imo. What do you think?

[Sly Fox]

Heh some guy on Beyond3d Forum have said that GT200 is THREE times faster than GF8800GTX in 3D Mark Vantage Could it be true? I`d expect something about 2 times faster maybe 2,5 in some situations but 3 times? It`s impossible imo. What do you think?

I certainly can do no more myself other than speculate, but if that's true Nvidia can count on receiving my money just as soon as they'll accept it

I've been waiting a long time for a G80 "killer," and I really do hope it comes out sooner rather than later.

[Scimitar]

Really?
This is why it makes me so hard to believe anything you say when you say hilarious things like this.
AA is done on the shaders. 480SPs clocked at +900mhz will be fine.
16ROPs clocked at +900mhz w/ tweaks will also be fine.
Please remember what G92 is doing with "only" 16ROPs...

Please provide some proof that AA is being done in shaders on the RV770. A well informed friend of mine told me AA is being done in ROPs with this new chip. I'm not saying you're wrong, but what is the source of your info?

Thanks.

[BenchZowner]

AA is done on the shaders. 480SPs clocked at +900mhz will be fine.

I wouldn't reply to this, as it's not my game to play the stubborn guy over disagreements but... 64 SPs vs 96 SPs ( the so called "320" & "480" SPs ) isn't that much of a difference.
I've run games with the HD3870 @ 999MHz and still it sucked b***s with AA & AF enabled.

I hope this time they got it right

[Barys]

VR-Zone have said the have got some info about Rv7700xt performance which is about 2 times better than Radeon 2900Pro in 3D Mark06 (9k vs 18k). If it is true then it seems that ATIs Midrange GPU Rv770 will be as fast as NVIDIAs Highend solution??

[shiznit93]

VR-Zone have said the have got some info about Rv7700xt performance which is about 2 times better than Radeon 2900Pro in 3D Mark06 (9k vs 18k). If it is true then it seems that ATIs Midrange GPU Rv770 will be as fast as NVIDIAs Highend solution??

link?

[Jowy Atreides]

link?

I can't remember the link but here's the pic I saved from it

http://userdisk.webry.biglobe.ne.jp/...c8f3d3857e.jpg

[shiznit93]

forbidden error.

[Barys]

link?

There is no link with 3DMark pic buy only info.

http://forums.vr-zone.com/showthread.php?t=267700

[NH|Delph1]

There is no link with 3DMark pic buy only info.

http://forums.vr-zone.com/showthread.php?t=267700

Not to be a killjoy, but this is fake.

In times where we want more facts, or just a clarifications whether it's Western or Eastern rumors that are closest to the trust, people have started to post obviously fake benchmarks. We're starting to get used to this by now, and sometimes the attention seekers put a lot of work into the forgeries, but not this time. At Chinese forum zol.com.cn, a user has posted a screenshot with GPU-Z where the name says "ATI Radeon HD 4870", cores says RV770 and node says 55nm. So far so good.

Then you look at the die-size; 153mm², and the ID; 1002-9589. These are for Radeon HD 2600XT (RV630). It doesn't stop there though. The next screen shows a finished run with 3DMark 06; 18,297 points, with an AMD Athlon 4400+ overclocked to 3,186MHz...

If we take the SM2.0, SM3.0 and CPU score and calculate the score we get 15,017. And then we have the fact that he used the wrong fonts for the stuff he pasted into the picture.

In any case, we just wanted to let you know that this is just a bunch of crap.

Source: NordicHardware



//Andreas