Of SKUs, OCs and what constitutes a midrange part

[DuckThor Evil]

At 2x to 3x the price the markets aren't even near crossing each other so there is just no reason to hold it back.

The markets are crossing each other quite heavily... I guarantee you that most Titan owners used to have 680s in their rigs. Ever heard about Tesla? That's a reason to hold the cheaper consumer versions back. They were making these chips last year and could have launched them last year as Geforce, but it made more sense to sell them as Teslas.

It's almost like saying that Intel can't release a 8 core Sandy-Bridge-E, but they easily could.

 

[silent_guy]

You'd think that silicon qualification for a Tesla class product is a bit more stringent than that for a consumer product. If they released the Titan as late as they did, it's because they wanted to.

 

[UniversalTruth]

Mid-range is generally considered to be from 200 to 293mm2. Everything above is high end.

Why exactly 293? I have never seen any evidence for what you just claimed. A link?

 

[Alexko]

I'm fairly sure that was a joke.

 

[DuckThor Evil]

wow the hits just keep getting better

 

[silent_guy]

Why exactly 293? I have never seen any evidence for what you just claimed. A link?

I thought it was interesting (well, not) how you picked a number of 300mm2. I took liberty to pick my own random number. 2.5% more or less, what difference does it make, right?

The fact that the first google hit for 'gk104 die size' results in 294mm2 is nothing but a coincidence.

Not that it is interesting or matters: 'high-end' is product positioning after taking all factors into account. Die size has nothing to do with it. After all, g92 came into market as high-end and then migrated to mid-range. I expect the same will happen with gk104.

 

[UniversalTruth]

I don't believe in artificial product positioning. It doesn't matter what they say but what I see. Titan is the proof how shitty "high-end" GK104 is. GK104 is a middle-class product by all sane standards. Period

 

[jimbo75]

Any business will take the opportunity when it arises. The first iteration of Tahiti was a joke that should never have seen the light of day, and AMD deserved to be beaten by GK104. They lost mindshare, marketshare and everything else deservedly and I doubt they'll make the same mistake again.

Now we see GK104 for what it really is, but it's not Nvidia's fault for taking advantage of AMD's lack of performance. Had AMD turned up at 28nm with their A-game it would never have happened.

 

[Silent_Buddha]

I don't believe in artificial product positioning. It doesn't matter what they say but what I see. Titan is the proof how shitty "high-end" GK104 is. GK104 is a middle-class product by all sane standards. Period

Really? And what would you say if you knew absolutely nothing about the die size? Only thing you knew that it was as fast as the competition, and that it was the best thing the company could make at the time of release because they were not capable of releasing anything faster.

Because that's really all most consumers that might be buying a card is going to know or need to know.

Reviews going into die sizes is a relatively recent thing. Back when 3dfx were battling it out with Nvidia pretty much no reviews went into die sizes because it is irrelevant to the general consumer.

The only thing the die size matters in is how it may impact the company's bottom line. Big die + low price (4870 versus GTX 2xx price war for example) = disasterous consequences to all involved. But the consumer doesn't need to know about any of that. Unless it makes a company lose so much money that they go into bankruptcy. But before that happened, both Nvidia and AMD decided that a price war wasn't good for either of them.

So really, the fact that GK104 didn't have a 500 mm^2 die doesn't stop it from being the top tier product that Nvidia had to offer based on what they could manufacture at the time and how it performed versus the competition.

Regards,
SB

 

[UniversalTruth]

Really?

Yes, really

what would you say if you knew absolutely nothing about the die size?

I looked at the performance charts and see that 680 performs very bad under given scenarios which gives a clue that it is the middle class product which was just rebadged in order to screw customers more, using of course the competition's weaknesses. Which by the way they show all the time

the company could make at the time of release because they were not capable of releasing anything faster

I have no information for this. I just know how pleasantly surprised NV was when they saw Tahiti's performance. So, could be that GK110 could have made it just a little bit later on the market but , for sure, it is ridiculous to claim that GK104 is high end just because it was released first