NVIDIA to Counter Radeon HD 6970 ''Cayman'' with GeForce GTX 580

[Sn0wm@n]

right ... so close to the amd 6000 launch this appears

[Bobsama]

Where did you get your numbers? Pretty much every review I've seen shows the 480 exceeding the 5970, in both average and maximum power. See here for example.

I'm saying what the max draw is for the finished boards. 99._% of finished cards will have lower draw than the max TDP.

[Lanek]

You use the numbers from review we have seen this years... but they use strange numbers anyway... (they get numbers for estimating their maximum power draw in watts, then soustract the system they use, and then they arrive to the what need the gpu ..) ... and ofc this is nothing to do with TDP ) i still want to know how they have determine the power ask of their I7980x OC@ 4ghz + rest of the system for say, the gpu power draw is x - y = Z )

[Sn0wm@n]

Thats gonna be hard to produce..... Looks good on paper tho.

everything can be made to look good on paper ...

Current fud says they're just keeping it as 5770 without releasing a 6770.

fud as you clearly said it ... wait until amd release real info about the 6k lineup so you wont make those fud statement anymore ...

That and the HPC market (Tesla). I don't even think AMD has an HPC product.

LOL so what are the firegl products???? clearly a fanboy statement ...

Well the 5970's total power draw is pretty close to 300W max. The GTX480, on the other hand, is about 250W max. For a single card with a single GPU, 250W is insane.

dual gpu card vs single gpu card ... what a nice comment ....

[Carfax]

LOL so what are the firegl products???? clearly a fanboy statement ...

Not quite. Firegl is a workstation product.

According to Bobsama they have something called Firestream though, which is geared towards HPC so I was wrong.

Here's a pic and some specs..

It looks relatively feeble compared to Fermi however, seeing as it's still based on the R700 architecture.

[Sn0wm@n]

amd firepro cards and firegl cards ... :P yes they do have ati 5k series gpu ... do a proper search mate ...


http://www.amd.com/us/products/works...s/v9800.aspx#4 1600 shaders ... ohh that sounds like a 5870 ... :O 4GB frame buffer ... etc... so sure ati sucks nvidia owns all the pro market ...

[Bobsama]

Firestream 9350 is HD5850, 9370 is HD5870. As for HPC performance, performance depends greatly on the code's optimization. For example, AMD graphics cards dominate in Milkyway@home whereas nVidia graphics cards lag (far) behind.

[Jokester_wild]

dual gpu card vs single gpu card ... what a nice comment ....

Might want to re-read what his statement said genius.

[Sn0wm@n]

Might want to re-read what his statement said genius.

the part about the one needing more juice???

[blindbox]

the part about the one needing more juice???

He said 250W for single card single GPU is insane.

[Sn0wm@n]

yes he did... and my comment was ... what a nice comment... in a sense to show that he did say a nice comment ... i could have said what a fanboy'ish thing to say if i were to disagree with what he said .. but i didnt disagree hence .. my approval ... wasnt it obvious in the first place ????

[Sam_oslo]

Lets hope nVidia can drop something (whatever that can push performance up or prices down), at the same time as HD69xx. But this time it should be at the same time, it will be good for us frequent consumers of GPU.

But I'm afraid it won't happen, these guys seams to have come to an unwritten-agreement to wait 6 months to counter each other though. At least they did it in the whole of last year. It's debatable if they are avoiding to compete directly (for keeping the pieces up in both camp), or it's just an innocent glitch.

[Sn0wm@n]

Lets hope nVidia can drop something (whatever that can push performance up or prices down), at the same time as HD69xx. But this time it should be at the same time, it will be good for us frequent consumers of GPU.

These guys seams to have come to an unwritten-agreement to wait 6 months to counter each other though. At least they did it in the whole of last year.

so you are suggesting that two different company made a deal to wait a said amount of time before they release something new for no apparent reason??? ... that would be totally against competitive marketing rules ... + against all logic rules ....


nvidia wants your money .. so does amd ... and both have different roadmap and if one stumbles because they were too much ambitious its their problem ... it doesnt have anything to do with a silent agreement ...

[Sam_oslo]

so you are suggesting that two different company made a deal to wait a said amount of time before they release something new for no apparent reason??? ... that would be totally against competitive marketing rules ... + against all logic rules ....


nvidia wants your money .. so does amd ... and both have different roadmap and if one stumbles because they were too much ambitious its their problem ... it doesnt have anything to do with a silent agreement ...

I don't know really, but something fishy seams to be going on in the whole last year. Neither of them has counted the other one at the right time.

We have only tow major GPU-makers, and they can either counter each other (at the same time) and push the prices down, or wait 6 months. It's debatable and I'm just wondering why they have avoided to compete directly.

This round can cast some light on the matter tho, lets see if nVidia will do something at the right time.

[Zloyd]

Dunno it just seems to me that nvidia are playing catch up doesnt seem to be anything else , while amd is progressing as per their schedule. The complications with fermi have probably just made things a little harder for nvidia so thats slowing them down as well.

[Sn0wm@n]

that something fishy comes from your imagination mate

[tajoh111]

that something fishy comes from your imagination mate

They were in cahoots and were found guilty of price fixing a couple years ago.

[Sam_oslo]

Dunno it just seems to me that nvidia are playing catch up doesnt seem to be anything else , while amd is progressing as per their schedule. The complications with fermi have probably just made things a little harder for nvidia so thats slowing them down as well.

Yeah this seams to be the start of it. But then ATi did nothing to counter GF104 either. Not only didn't release a refresh, but also no price-drop to compete directly. It was surprising, ATi did nothing, nada.

As a result, we didn't get a direct competition, and we got the highest GPU-prices (in both camps) in the whole last year.

So I'm wondering if they will ever go back to good old days and counter each other at the right time. Why should they after a year with study high prices? It will be interesting to see if the trend will continue in this round too.

[SimBy]

Will this puppy be a real product or a mock-up with wood screws?

[Sn0wm@n]

They were in cahoots and were found guilty of price fixing a couple years ago.

huh??? proof ????

[SimBy]

huh??? proof ????

It's true. Too lazy to google for article.

[spursindonesia]

I agree. If you're fooled twice by this, you're the fool. Nobody should be paying attention to whispers in the wind about new nvidia products. Either you talk about something real, or you're full of

Yep, talk is cheap. Let's see a FULL 512 SP GF100 card sold in good quantity before talking about next gen flagship chip from the green corner.

512-bit memory controller is way too high.

576SP/96TMU/384-bit would be nice

This is also my bet as nVidia's 40 nm response in late Q1/early Q2 2011 against Northern Islands's big gun, Cayman. Ofcourse, Antilles will continues Hemlock's fate as unchallenged king of the hill card of its lifetime.

GF110 mit 768SE neuer Chip mit 768 Shader-Einheiten und 128 TMUs an einem 512 Bit DDR Speicherinterface, ohne den GF100-Ballast nach GF104-Vorbild, Die-Fläche mit 550-650mm² größer als GF100 und damit auch höhere Verlustleistung, Zielmarkt dieses Grafikchips wären dann eher die DualChip-Lösungen von AMD
Potential gegenüber GeForce GTX 480: ca. +50%



Chances of that happening on 40nm though?

But holy hell, 768 shader, 128 TMU, 512bit with 2GB? Would love something with that kind of performance in single GPU.

Find it a little annoying that articles like the OP's include Dual GPU cards when talking about "performance king". That kind of stuff needs an asterisk. Single GPU power should be where the real credit is given, not that I'm a fan of NVidia.

That specs isn't going to happen in 40 nm node as a single chip, while as a dual chip card, only feasible if AMD decides that Antilles is a dual Barts card rather than dual Caymans, but i think the chance is pretty darn close to none.

Maybe, maybe not .. but let's face it... if all the series bring new feature as 3D, UVD3, new display features and other things, if they use the 5770 it will appears to be the low end, inititally dedicated for movies, HTPC (6400?).. you will get lower gpu ( Caicos, Turk ) with new features, then a gap with 5770 with not thoses features, then 6870 and higher end model with thoses features ? ...

maybe they will keep 5770 and then replace them in the 6K series later, but i really don't think they can bring new features on really low class gpu ( caicos ), average ( Barts ), high end ( cayman ), and keep a little 5770 ( renamed or not ) with the old features between 6400 and 6870... Ofc all the features i talk about are side of the gpu itself, maybe they will use the Juniper core and add just the 3D, UVD3 engines, and new displayport, HDMI on the PCB and then rename them 6450-6650 ....

My opinion, HD 57xx will stay until replaced by 28 nm pilot chip cards in late Q2 2011, Turks would be a HF 66xx while Caicos, i don't have a clue about this chip fate.

[kgk]

As long as Cayman has good minimum frame rates, that's all I care about. Haven't bothered with 5000/GF100 cards (waiting for next gen) but from what I recall isn't Fermi much better than the 5000 series at minimum frame rates at identical resolutions/settings?

I would much, much rather have a card that runs something at 1920x1200, max details @ 50FPS average with 35FPS minimums than something that does it @ 80FPS average with 20FPS minimums.

[rmanuelb]

as long as cayman has good minimum frame rates, that's all i care about. Haven't bothered with 5000/gf100 cards (waiting for next gen) but from what i recall isn't fermi much better than the 5000 series at minimum frame rates at identical resolutions/settings?

I would much, much rather have a card that runs something at 1920x1200, max details @ 50fps average with 35fps minimums than something that does it @ 80fps average with 20fps minimums.

x2 !

[informal]

With Antilles coming out in Dec,Nvidia will need a GPU miracle to regain perf. crown.Not that they have it now,since 5970 is faster than GTX480,but they will loose single-GPU perf. crown to Cayman in a month.IMO even if they double everything in their next gen GPU(compared to GF100),the card based on that GPU would still probably loose to Antilles all around.And by that time,Antilles will be old news,since next iteration of Radeon is coming out,with better performance at the same price points(as it was the case in the past).