NVIDIA GeForce GTX 980 Ti

[CarstenS]

Ok, let's put it the other way round. If all those points were in AMD's product favour, do you honestly believe that the market would hug it ?

I think NO, it is a predetermined choice, if you wish even on psychological level. Maybe those favours could sometime put some slight direction but do not change the whole picture.

I see it thus: A cooler and memory technology are means to an end, a cooler being more of a crutch anyway.
Both have to serve the GPU in a way, that it can unleash the best performance for my games. They need to be as unobtrusive as possible. In case of the cooler, it has to be easy enough to handle for installation and it has to do it's job as quiet as possible. In case of memory it has to keep the GPU fed with data.

In both cases, I as a consumer am not interested in nice new buzwwords but in the effect it has on my investment. As a technology enthusiast, I applaud AMD for introducing HBM, but that won't pay their bills in the big picture.

 

[silent_guy]

You forget that currently R9 295X2 is the fastest and most attractive cheap solution. Small die produce excellent dual-GPU cards.

Sure, if you can about a dual-GPU space heater.

 

[Deleted member 13524]

You forget that currently R9 295X2 is the fastest and most attractive cheap solution. Small die produce excellent dual-GPU cards.

Yap. It's still impressive to see how the 295X2 gobbles up each and every benchmark at 4K.
If one has the power supply to handle it, dual Hawaii seems like the very best value to get in the $650-700 zone, granted that you'll have to wait for a driver to support each high-profile games.

 

[Albuquerque]

Since someone asked, I too waited for the FuryX reveiws to come out before making my purchase.

My video 3D card history goes something like Riva128 -> Matrox MGA-G200 + Voodoo 4mb -> GeForce 3 Ti200 -> Radeon 9500np (modded) -> Radeon 9800np -> Nvidia 7950 -> Radeon 3850 CF -> Radeon 4850 CF -> Radeon 5850 -> Radeon 7970 OC -> ... (n+1)

(Later edit: I'm forgetting a few, there was an x800xt in there somewhere, and an 1850Pro or something)

All things considered, ATI / AMD have collected quite a bit of my money over the years. I would also say that I made these decisions not because I'm an ATI / AMD fanboy, but because every video card I bought had the performance I was seeking, the features I wanted, and a price point I was willing to pay.

This is the first time in a long time, for me, that AMD simply didn't have the compelling argument against NV. For how I intend to use my card, my Gigabyte 980Ti G1 delivers everything I want at a price I was willing to pay.

 

[UniversalTruth]

Technically speaking Fury X is a better product than 7970. If you put them next to the competition. Back then 7970 was even slower than 680. Now you have the Fury X slightly beating in 4K the 980Ti.
4 Fury X completely demolish a quad slid 980Ti. CF works better than Sli.

So, I do not understand you.

 

[Razor1]

hmm he wasn't interested in the CX, CX/SLi is a very limited market.

And when we compare single cards, the 980 TI is just better at the same price, with its lower power consumption and ability to overclock. Its not always just about performance, its about everything the card can offer at a certain price.

The 480 gtx did not sell well as the 5xxx series. nV lost marketshare even though they had the fastest GPU on the market but at a cost, 100 watts more under load. They didn't re-stabilize and gain marketshare until their midrange cards hit.

 

[RecessionCone]

You forget that currently R9 295X2 is the fastest and most attractive cheap solution. Small die produce excellent dual-GPU cards.

Hawaii is not a small die.

 

[Albuquerque]

In a "why did you purchase 'X' computer component" conversation, I still do not understand why die size even matters. I understand why it's interesting technologically, I also understand why perf/mm can be a itch that your curiosity wants to scratch upon. As a consumer, I couldn't give two lumps about die size (and any mathematical function that uses die size in the equation.)

I care about price, performance, and features. If two competing cards have die sizes of 3mm^2 and 300mm^2, and yet produce the same performance and features for the same price, then as a consumer they're equivalent. The die size, as a function of resulting capabilities, doesn't matter to a consumer.

If the larger die costs more, then the cost is the conversation -- not the die size. As a consumer, I shouldn't care about why your cost is more, only that it is more and that the performance or features must (according to my own biases) account for that.

 

[RecessionCone]

In a "why did you purchase 'X' computer component" conversation, I still do not understand why die size even matters. I understand why it's interesting technologically, I also understand why perf/mm can be a itch that your curiosity wants to scratch upon. As a consumer, I couldn't give two lumps about die size (and any mathematical function that uses die size in the equation.)

I care about price, performance, and features. If two competing cards have die sizes of 3mm^2 and 300mm^2, and yet produce the same performance and features for the same price, then as a consumer they're equivalent. The die size, as a function of resulting capabilities, doesn't matter to a consumer.

If the larger die costs more, then the cost is the conversation -- not the die size. As a consumer, I shouldn't care about why your cost is more, only that it is more and that the performance or features must (according to my own biases) account for that.

The various execution gloom threads show why it matters to a consumer - low margins lead to low R&D investment, which leads to weaker products, which lead to lower margins, etc.

 

[Albuquerque]

The various execution gloom threads show why it matters to a consumer - low margins lead to low R&D investment, which leads to weaker products, which lead to lower margins, etc.

Nope, as a consumer, I couldn't care less. As a patron of Beyond3D, yes I am at least interested.

If a card with a larger die gives me the same performance, features and price as a card with a smaller die, they're equivalent. Next year, if the R&D budget went down as you describe, then the successive product will have less performance, or less features, and therefore better show up with a price that is commensurate with the reduced capability. If my needs as a consumer cannot be met in the face of that performance and/or feature loss, then the price simply will not matter anyway.

Die size doesn't matter for a consumer.

 

[RecessionCone]

Nope, as a consumer, I couldn't care less. As a patron of Beyond3D, yes I am at least interested.

If a card with a larger die gives me the same performance, features and price as a card with a smaller die, they're equivalent. Next year, if the R&D budget went down as you describe, then the successive product will have less performance, or less features, and therefore better show up with a price that is commensurate with the reduced capability. If my needs as a consumer cannot be met in the face of that performance and/or feature loss, then the price simply will not matter anyway.

Die size doesn't matter for a consumer.

I agree - the impact of bigger die size on the consumer is in the murky future.

 

[CarstenS]

Technically speaking Fury X is a better product than 7970. If you put them next to the competition. Back then 7970 was even slower than 680. Now you have the Fury X slightly beating in 4K the 980Ti.
4 Fury X completely demolish a quad slid 980Ti. CF works better than Sli.

So, I do not understand you.

Back then, when HD7970 came out, there was no GTX 680.
Fast forward today, when 980 Ti came out, there was no Fury X.

 

[Albuquerque]

Back then, when HD7970 came out, there was no GTX 680.
Fast forward today, when 980 Ti came out, there was no Fury X.

This gets into grey area in my head... As a pure consumer, a few weeks head start on people blithely ordering whatever is the fastest might account for some measurable quantity of sales. However, at the price that we're talking about, I suspect most of the target consumers will have at least some knowledge that AMD has a product right behind...

There is also the case of those who are fiercely loyal to their brand, so the new NVIDIA drop would invariably result in a slew of fanatics buying the newest and best. Still, they wouldn't have bought the AMD product anyway, so the timing variance seems like a non-sequitor in this instance.

Purely as an anecdote: I waited for the GTX 680 to land before I bought the 7970, which resulted in the 7970GHz edition then being pumped out. All things considered, I feel that a consumer of this "level" of purchase would have done similar research as I did. I also understand that my own biases will color my judgement of others

 

[fbomber]

Nope, as a consumer, I couldn't care less. As a patron of Beyond3D, yes I am at least interested.

If a card with a larger die gives me the same performance, features and price as a card with a smaller die, they're equivalent. Next year, if the R&D budget went down as you describe, then the successive product will have less performance, or less features, and therefore better show up with a price that is commensurate with the reduced capability. If my needs as a consumer cannot be met in the face of that performance and/or feature loss, then the price simply will not matter anyway.

Die size doesn't matter for a consumer.

Driver development can be an issue, if the company has less money.

 

[Razor1]

Driver development can be an issue, if the company has less money.

That is true but wouldn't driver development be spread across R&D and and regular expenses?

 

[fbomber]

That is true but wouldn't driver development be spread across R&D and and regular expenses?

Even if that´s the case, the R&D budget of a struggling company is lower, no?

 

[Razor1]

Even if that´s the case, the R&D budget of a struggling company is lower, no?

True, I agree with you, just wondering how it would be split up that's all

 

[Rys]

I've cleaned the back end of this thread up so it remains on topic. If you have something to say about another member with no other on-topic content to add, there are other means to communicate that. Don't do it in product analysis threads.

 

[Albuquerque]

Driver development can be an issue, if the company has less money.

A consumer doesn't have a way to see this. Would you expect the company's last three financial filings to be posted on the box, in an effort to show that they'll have better support in the future? How did that work out for NVIDIA's drivers back when Vista first came out, when they were the single largest logged source of BSOD's on the new OS? Howabout Intels' brazillion dollars that they do not spend on GPU driver updates for integrated graphics that are more than two generations behind?

Just like a stock investment strategy, past performance does not guarantee future results. There is no credible correlation of corporate margins to driver quality.

I remain unconvinced that die size has any relevance to consumer purchasing strategy.

 

[ninelven]

Err... Intel actually spends a fair amount on driver development AFAIK. Maybe not as much % as their gross income, but that isn't really a fair comparison.