Nvidia just posted this on Twitter: This is TITAN X

[Pomfinator]

It's only a rumor to be fair, but given the enormous 630+mm sq. die, it's possible yields are in the toilet.

If that's the case, I would be a whole lot less excited.  I wouldn't even be mad if they just charged the old Titan price for it

[Briggsy]

That Price... If money was no object, 980 SLI would get crushed to death (until 6GB 980Ti).

 

[Opcode]

Can't you fold using your phone these days?

I think he got confused of the context and think we were talking about TITAN X vs Vesuvius when we were discussion the Shield being a hotter piece of hardware than the TITAN X because of its price point being more acceptable to everyone in the market. Tho yeah, if you wanted to fold on your Shield I imagine there's not much stopping you from doing so.

[patrickjp93]

Who's going to use their Nvidia Shield for compute?  :huh:

 

Shield = $199

TITAN X = $1400

 

Not everyone can afford the TITAN X tho the Shield is within reach of many more consumers.

 

Too bad you get better performance/price density with the Titan X. 7*256 = 1802 CUDA cores vs the 3072 on the Titan X.

[Faa]

No I am saying its simply better value than the titan Z.

Not even a single guy on the planet has been trying to say the Titan-Z is a better value than the 295x2..

You're repeating the same old story people just don't want to hear again.

[Opcode]

Too bad you get better performance/price density with the Titan X. 7*256 = 1802 CUDA cores vs the 3072 on the Titan X.

I wouldn't be buying a Shield for its compute performance. I'd be buying it to use it for its intended purpose.

[patrickjp93]

No I am saying its simply better value than the titan Z.

For you, this may hold true. It's not true for everyone. If you have HPC programs implemented in CUDA, or if you need a card which can do both OpenCL and CUDA for direct comparison, then Nvidia is pretty much your only choice. Value is composed of different things for different consumers.

[Carde]

Now sell it for 500 USD/EUR and I be interested.

[othertomperson]

No I am saying its simply better value than the titan Z.

 

But it isn't. They are literally the same value. The proportional markup of two individual GPUs over Nvidia/AMD's amalgamation was pretty damned identical. Stop pretending that AMD doing the exact same thing as Nvidia was any better.

[spartaman64]

Put a Titan Z under water and it's the 295x2 that gets smashed to be fair.

not with how much the titan z is underclocked

[Guest]

Not buying those games is always an option, or are you willing to give up your power as a consumer and pay more than you should have to for a given experience?

 

Not wanting to face reality is just the approach of the mentally immature.

 

The whole point is to let a budding programmer or startup studio do it cheaply (meaning you go without the insurance of ECC). It serves a niche but distinct and important purpose: lowering startup costs.

 

Instead of Teslas my school just tossed a pair of titan black cards into one of our server blades for the HPC and Heterogeneous Computing classes to use. That way we could experiment on a variety of program scales but not have to pay 3-4x more for the equivalent Quadro or Tesla cards.

 

Believe it or not if you dig around in the driver code the professional and geforce drivers are about 95% similar. Software does not rely on professional drivers. It only assumes they are there and are tuned to perform based on the existence of those drivers. What actually happens on the hardware is transparent to the program as long as outputs are correct.

 

Nvidia GIMPs to be able to run cooler, hence its popularity with ITX system builders.

 

the argument is not silly at all. Only your counterargument is.

 

Of course that's an option, but why would you let that restrict you if the rest of the game is 100% fine? "They use more VRAM than they actually need to" is not the end all and be all criterium for the value of a game. It's actually a really silly criterium and something that can easily be compensated for by the GPU. Of course it's not ideal to waste resources, but this is reality after all and it's going to happen to a lesser or greater extent one way or another.

 

Now that's a huge difference, helping out startups and schools with a more affordable option is one thing, but that does not equal a "workstation card". Doing it without ECC means doing it wrong, if it's your only option (because you can't afford a proper workstation card) or if it's for a school (where mistakes don't matter that much) then it's obviously a good choice (and as you said, a niche), but that still doesn't make it a workstation card that could be used on time- and/or error-critical projects.

 

The fact that the drivers are that similar is interesting, I didn't know that. But even so, it lacks some of the features that are exclusive to Quadro GPUs and that may or may not be important for what you're doing with it.

 

 

My argument still stands from a consumer perspective, the Titan line-up is harmful and drags down the GeForce cards one way or another. Because fact is, many people buy them as pure gaming cards. And that's the reason why we didn't have a 6GB 780ti. If they truly didn't compete and the Titan was a "developer card" then there would have been no reason to do that, especially since we had a 6GB 780 (because that doesn't directly compete with any of the Titans or a 780ti). What other reason would there be?

And if what you're saying is true and the Titan X doesn't even feature proper DP functionality anymore, then that's even worse. That literally makes it the "you get it early" and "it has more VRAM" version of a normal GeForce GPU. Sure, people will buy it and they have every right to do so, but it's also my right to point out that this is fucking retarded and harmful.

[Arct1c0n]

Costs as much as my entire rig basically. Sorry, I just don't have that kind of disposable income for e-peen. But this will at least shut up all the memory whiners

[Opcode]

For you, this may hold true. It's not true for everyone. If you have HPC programs implemented in CUDA, or if you need a card which can do both OpenCL and CUDA for direct comparison, then Nvidia is pretty much your only choice. Value is composed of different things for different consumers.

The smart thing to do would be to convert the kernels over for OpenCL support and run Vesuvius.

[spartaman64]

It only wins because AMD put it under water. If you put the titan Z under an EK waterblock, guess which one actually wins in performance (and still in perf/watt)...

 

I stand by what I said AMD produced a failure in engineering and had Asetek bail them out. All things being equal the Titan Z can win, though its launch price was obnoxious and I'm not going to deny that.

lol when liquid cooling is considered a bad thing

[Pomfinator]

Costs as much as my entire rig basically. Sorry, I just don't have that kind of disposable income for e-peen. But this will at least shut up all the memory whiners

You know you want it, that sexy shroud is calling for you.

[Meladath]

lol when liquid cooling is considered a bad thing

 

Only because its AMD doing it.

 

The day nvidia releases a GPU with liquid cooling by default it will be praised as the most intuitive innovative best thing ever and people won't stop shutting up about how amazing it is.

[patrickjp93]

not with how much the titan z is underclocked

You're missing the point. The 295x2 is greatly overclocked due to being underwater. Under water the Titan Z can be pushed to 1100+ MHz. At that point the 295x2 gets buried.

 

[spartaman64]

You're missing the point. The 295x2 is greatly overclocked due to being underwater. Under water the Titan Z can be pushed to 1100+ MHz. At that point the 295x2 gets buried.

but nvidia didnt take that path and they decided to nerf their clocks and if they did do that the price would be probably 4000

[digitalnav]

Last time the dual chip battles happened the 295x2 won over the titan Z easily.

Considering a price/performance standpoint or just a performance standpoint in SOME situations.

And that is just a fact.

Hey look, its my video!

[patrickjp93]

Only because its AMD doing it.

 

The day nvidia releases a GPU with liquid cooling by default it will be praised as the most intuitive innovative best thing ever and people won't stop shutting up about how amazing it is.

 

Nope. It's an engineering failure to release a REFERENCE version as liquid cooled (especially on a locked-down loop). Offering it as the premium version is one thing. Being forced to rely on liquid cooling to be able to get the card to manageable heat levels is a failure of engineering, regardless of the company.

[MrRavens]

I want the benchmarks...now!

[Arct1c0n]

You know you want it, that sexy shroud is calling for you.

 

Refrence blowers suck for cooling, no thanks.

[Fhyrion]

So it's probably gonna be the most powerful consumer grade single graphics card available, and it will also have a terrible price to performance ratio. 

In other words, it's another titan. What did anyone really expect.

[Meladath]

You're missing the point. The 295x2 is greatly overclocked due to being underwater. Under water the Titan Z can be pushed to 1100+ MHz. At that point the 295x2 gets buried.

 

"Greatly overclocked"

 

1018Mhz is greatly overclocked? A whole 18 Mhz higher than an R9 290X?

 

Wow, what a HUGE overclock!

 

Because the R9 295X2 can't overclock any more at all right?

[mxthunder]

Hope it stomps the 390X into the ground