Now The Real Battle...... START

[dead_95]

Its shows that Nvidia next generation GPU .............code name:fermi.........product series: Geforce 100 series>>>looks very

promising............

Will it demonish ATi (am a Nvidia fan..) lastest and oncoming GPU from the red team.............

only the future will tell us...


<RED> .............OR..........<GREEN> ........

<WHO WILL WIN>>..??

 

[mfarrukh]

NONE

Now it's ATI

TOMMOROW will be NVIDIA

THE DAY AFTER TOMMOROW will be ATI

THE DAY AFTER that will be NVIDIA


This would never end

 

[sayantan]

+5^^whatever ati does ,nvidia will come up with an answer and vice versa

 

[Truhls]

The chances of the Fermi bringing nVidia over ATI are very slim so far.

http://www.semiaccurate.com/2010/01/17/nvidia-gf100-takes-280w-and-unmanufacturable/

EVEN if it does do better then ATI's best, it doesnt matter. it costs 2.5x as much to make as the ATI equiv, so there is no way it can be priced competitively without nVidia taking a loss.

 

[mfarrukh]

^^^^^^^

HA

ATI FAN

 

[Truhls]

^^^^^^^

HA

ATI FAN

No actually, ive only ever used nVidia GFX cards so far ( though i will be buying ATI here shortly ). im just not a blinded idiot fanboi to think nvidia is going to be like OMG WTF I MUST HAVE DIF CARDF!!!.1.111oneoneone.

 

[hallowed_dragon]

We will have to see because for the moment we only have a cherry picked benchmark that if put in the normal reviewer perspective is a bit underwhelming.

 

[mfarrukh]

No actually, ive only ever used nVidia GFX cards so far ( though i will be buying ATI here shortly ). im just not a blinded idiot fanboi to think nvidia is going to be like OMG WTF I MUST HAVE DIF CARDF!!!.1.111oneoneone.

Yea, But NVIDIA know market well.
The are aware of the current market situation of how ATI is dominating for couple of months.


So, you can't go on saying ''EVEN if it does do better then ATI's best, it doesnt matter. it costs 2.5x as much to make as the ATI equiv, so there is no way it can be priced competitively without nVidia taking a loss.''


And how do you know that FERMI costs 2.5x greater?


Even That article is not real.

Tell me, that the guy who wrote it has made the FERMI card.

 

[Truhls]

Yea, But NVIDIA know market well.
The are aware of the current market situation of how ATI is dominating for couple of months.


So, you can't go on saying ''EVEN if it does do better then ATI's best, it doesnt matter. it costs 2.5x as much to make as the ATI equiv, so there is no way it can be priced competitively without nVidia taking a loss.''


And how do you know that FERMI costs 2.5x greater?


Even That article is not real.

Tell me, that the guy who wrote it has made the FERMI card.

Maybe because it has been confirmed fermi i using a far larger die then the ATI models? And because of that the chips per wafer are going to be less, and any defects on said wafer are more significant. All of which i common sense for any idiots to understand. that alone makes the fermi more expensive to make. and that is something that has been PROVEN.

And the article probably isnt all real, but it probably has some truth to it and it is a good read besides.

Also with nVidia faking the fermi based GPU, that is just lawl.

 

[mfarrukh]

You seen it?

How's that proven.
You have the card?

WOW, you are a lucky man to get it before anyone else.


WOW

 

[hallowed_dragon]

Yea, But NVIDIA know market well.
The are aware of the current market situation of how ATI is dominating for couple of months.


So, you can't go on saying ''EVEN if it does do better then ATI's best, it doesnt matter. it costs 2.5x as much to make as the ATI equiv, so there is no way it can be priced competitively without nVidia taking a loss.''


And how do you know that FERMI costs 2.5x greater?


Even That article is not real.

Tell me, that the guy who wrote it has made the FERMI card.

Actually the Truhls is a right in some points, although in a fanboish way. I don't know if the cost of a Fermi chip is 2.5 times larger than Cypress, but it is a lot higher.

 

[Truhls]

You seen it?

How's that proven.
You have the card?

WOW, you are a lucky man to get it before anyone else.


WOW

I am fairly certain nvidia has stated themselves what the die size is going to be. and if you dont believe nvidia then why are you their diehard fanboi?

 

[mfarrukh]

HA

Who is the fanboy?

ME?

I am just telling you that you can't be certan that FERMI will cost this and run this hot and E.T.C.


For the record,
I never had a NVIDIA card off my 4 Video Cards.

 

[Truhls]

HA

Who is the fanboy?

ME?

I am just telling you that you can't be certan that FERMI will cost this and run this hot and E.T.C.


For the record,
I never had a NVIDIA card off my 4 Video Cards.

Regardless if what you say aside from what, a new respin ( which means waiting probably another 6 months for fermi ) no matter how you look at it fermi will cost more then ATI 's best chip to make. If nvidia decides to take the hit and have competitive pricing then good for them, either way its still a loss.

 

[sayantan]

The chances of the Fermi bringing nVidia over ATI are very slim so far.

http://www.semiaccurate.com/2010/01/17/nvidia-gf100-takes-280w-and-unmanufacturable/

EVEN if it does do better then ATI's best, it doesnt matter. it costs 2.5x as much to make as the ATI equiv, so there is no way it can be priced competitively without nVidia taking a loss.

Hey buddy ,don't underestimate nVidia,its a much larger company than ATI(not AMD-Ati) and will be pumping in a lot more cash on its R&D....

 

[mfarrukh]

Well, the same thing again

Make thousands of predictions, but you attitude seems ''certain''


I am just pointing out that you can't be certain about FERMI yet

 

[hallowed_dragon]

HA

Who is the fanboy?

ME?

I am just telling you that you can't be certan that FERMI will cost this and run this hot and E.T.C.


For the record,
I never had a NVIDIA card off my 4 Video Cards.

Actually from the great chip surface ~550 mm^2 (deduced by reliable sites) and the poor 40 nm yield of TMSC you can safely say that the Fermi chip will cost higher to produce. The unknown fact is the price with which nVidia will sell them to produce competition for ATI.

 

[renz496]

when Fermi and Cypress enter the rings another battle between Nvidiots and ATIdiots will go along with it

btw i'm looking forward to pure performance comparison between equivalent Fermi and Cypress regardless of price vs performance aspect.

 

[hallowed_dragon]

when Fermi and Cypress enter the rings another battle between Nvidiots and ATIdiots will go along with it

btw i'm looking forward to pure performance comparison between equivalent Fermi and Cypress regardless of price vs performance aspect.

I think we will have to wait for a little more. I hope that the Fermi will put nVidia back on track so I can get my ATI card cheaper

 

[notty22]

What I've read and understand this will at least do DX 11 faster and better looking. if that took a bigger die, then so be it. Its useless to take a 40% performance hit like ATI's cards do. Fermi was shown screaming through Unigenes Heaven dx11 benchmark. And they are making claims of 50% faster than 5870. What counts is what this will cost at the store front. Not what some angry ati fan boi' deduces from die size and other guesstistics.

 

[hallowed_dragon]

What I've read and understand this will at least do DX 11 faster and better looking. if that took a bigger die, then so be it. Its useless to take a 40% performance hit like ATI's cards do. Fermi was shown screaming through Unigenes Heaven dx11 benchmark. And they are making claims of 50% faster than 5870. What counts is what this will cost at the store front. Not what some angry ati fan boi' deduces from die size and other guesstistics.

You are falling for PR here mate. That is a synthetic benchmark and we all know what synthetic benchmarks mean in real world conditions. I believe you are falling for the hype instead of looking at the given numbers, which tell you something if you choose to listen. Just look at the youtube benchmark of a Fermi card against the 285GTX. We know they are real and we know the reviewer tested on Small ranch benchmark in Far Cry 2. We know the Fermi card scored 84 FPS. We alse know from reviews that the 5870 scores around that value + or - 2-3 FPS. Now the unknown here is the Fermi card tested. If it is the 360GTX then that puts it in the same ballpark as the 5870 and the 380GTX roughly 20% faster. That leaves Hemlock without competitor. If the card in question is the 380GTX I don't have to explain what that means.
That is only 1 benchmark and doesn't prove anything. Same goes for Uniengine.

My suggestion is to wait and see what will March give us. There is no point in starting flame wars for something that we have no data on.

 

[JAYDEEJOHN]

Just to get things back to earth here. Id remind everyone where these "benches" come from.
These are Fermi at their best. And hand selected at that. We still have to wait, as theyve really done nothing yet, just keeping their name in the race is all

 

[hallowed_dragon]

Just to get things back to earth here. Id remind everyone where these "benches" come from.
These are Fermi at their best. And hand selected at that. We still have to wait, as theyve really done nothing yet, just keeping their name in the race is all

You are right jaydee. The benches come directly from nVidia so they are cherry picked. As I said earlier, we can't make any assumptions regarding this data. We can only guess on the architecture and raw GPU power numbers which seem pretty nice.

 

[JAYDEEJOHN]

Im interested in their new approach to triangle setup, even did a thread awhile ago on it, but no one contributed.
This could be nice for tesselation, or not. From what Ive read, its close to the way ATIs shaders are layed out, meaning, yes 5/1, but those 5 arent often used, so too may nVidias solution regarding its triangle setup, and tesselation, tho time will tell, and benches arent real world, so unigine may have to be taken with a grain of salt, too early to tell yet

 

[hallowed_dragon]

Im interested in their new approach to triangle setup, even did a thread awhile ago on it, but no one contributed.
This could be nice for tesselation, or not. From what Ive read, its close to the way ATIs shaders are layed out, meaning, yes 5/1, but those 5 arent often used, so too may nVidias solution regarding its triangle setup, and tesselation, tho time will tell, and benches arent real world, so unigine may have to be taken with a grain of salt, too early to tell yet

From what I read regarding the Fermi tesselation(maybe I misunderstood) the chip is designed to use more shaders as the tesselation detailed is increased. ATI has a dedicated tesselation unit that doesn't change from 5750 to upper levels so I guess there is a limited power on the ATI side in this matter.