It's almost irrelevant in today market. How many do you think they will sell? The 970 and 980 are just way better value. To top it up, if AMD do a paper launch and it beats the TITAN at 2/3 of the price it will hurt Nvidia's sales even more.
I don't see it selling anywhere near as much as cards for your average Joe, I'm just pointing out it won't be DOA because of rumours. Until we see proper reviews and an actual product (390x) on sale, the TitanX will be a card for those that want the best now.
People complaining about the price of it... it's a freaking high end card for crying out loud. I remember paying $1100 for the X1900XTX on the week of it's release. You want high end? you pay high.
There will always be people buying the Titan because they want exactly that card. It isn't meant to be bought by the masses. Sure, 390X sales will always compete with the Titan, but the Titan is not meant to be bought by the people looking for a 700$ graphics card.
Exactly my thoughts. what is this? currently world's fastest single GPU? yeah you gotta pay for that "rank". Owning any top tier product is a luxury and for that you have to pay. I never thought i would pay 700 bucks for single gpu (780ti) but yet I paid, along with thousands around the world (hey i dont know the exact sales) just saying, if 700 is doable then 1000 is also doable.
Plus, the Titan series has always been this high in price range. What do people expect, a $500 gimped Titan?
Original Titan is still priced more than GTX980, thats just silly. I dont expect price droping anytime soon, even if 390X beats it for less.
Does Nvidia really care atm? They beat expectations with the sales from Maxwell. I doubt many of the people who have Maxwell now are going to upgrade to a 390x and even if they do, Nvidia already made their money. They couldn't have made a better card then Titan on Maxwell (28nm). The chip is 600mm2 which is absolutely insane. The only thing they could have done was clock it higher. Jens basically said that PASCAL will be due out by the end of this year, I suspect Q1 2016, but regardless that will bring HBM/16nm/2xP/W. 390x is now rumored for June release. So that means Nvidia will suffer with slightly slower, but more power efficient cards for what, 6-8 months before Pascal? In the meantime they carved out niche markets in both HPC/Cars and pushing over and over again to get into mobile. If the 390x's launch is indeed June I think AMD dropped the ball. They took a risk and bet on the right tech with HBM, but it's so late at this point that the advantage they would have gained has been somewhat squandered.
Why that, when you can have a 1000 dollar gimped Titan? The previous titans were entry level compute cards, this one isn't. This is just a video card. A thousand dollars is too much for just a video card. I mean maybe some people are fine with the price, but I don't see why you don't buy two 980s in 100% of cases. It's and faster, and essentially the same price.
Hm, I think that two(or more) of these beasts will make rich people (or crazy enthusiasts..) happier if they want to play on 4K with good fps. Two 980s can't. And that's all for new Titan..
there is a market where people will buy 2x titan X because they are faster than 2x 980gtx. this card is for that market. Your analysis is absolutely correct, 2x 980 will beat single titan X in price / performance ratio and then there are cards like 750ti which give you even higher price/performance ratio at lower price but you wont buy them because the performance is too slow for you. you tell this to a 750ti owner and he will say how much performance do you want? its already running at 60fps lol. what we see here is, as performance goes up, that ratio goes down.
As if AMD or Intel have never been tagging their high-end CPUs a thousand dollars. Exactly, some people are fine with it. Supply and demand; get over it. Some people also don't want the hassle of owning an SLI or Xfire setup, which never really work out of the box unless there's a proper profile for it. For them, a single-GPU card that performs about the same as a dual-card setup is worth not having to tinker with profiles and just play games.
While I do agree the architecture wise card is not much as a basic compute card as the original Titan was, Maxwell can perform and show the same behavior as a compute card. This still has more shaders, faster shaders, faster memory, and bandwidth available than what the 980 has. You can't really compare this to two 980's also. SLI isn't always perfect, not all games and titles have an SLI profile, compute applications that use OpenCL, CUDA, and other languages cannot take advantage of SLI, Linux and Mac do not utilize SLI, and if I'm correct there's still more power usage. I would compare this more to a Quadro than a Geforce though looking at it. People also buy the Titan series knowing the price tag for usually the same general reason, it is the fastest single card. Nvidia is not replacing their high end segment with the Titan, they will still release a high end card that is cheaper.
A quadro for what? Without the Quadro driver it won't even be accelerated properly in half those applications. Compute performance is limited to FP32 precision, which is useless in half decent math applications. Also compute applications can use multi-GPU, it's not technically "SLI" but they can distinguish both cards and share stuff between them. I mean I agree that there is a niche market for this product, but it's "worth" compared to previous Titan's (excluding Z) is simply not there. Especially when you consider the time it's launching and competitor rumors.
Well to be fair it definitely makes people question their decision. Like the rumors of the 390x alone are doing a good job at preventing people from making purchases. Obviously it doesn't help us as gamers but from a marketing strategy it's fairly effective. 390x should have been released a month ago though.
