Separate names with a comma.
Discussion in 'nVidia Flavor' started by Stoly, Mar 17, 2015.
Jen Hsung just said that for Double precision there's Titan Z
Or a Quadro.
That's interesting. That was pretty much the entire marketing point of the original Titan, selling it as an entry-level workstation card. Guess they're going to drop that? I don't see too many gamers spending $1399 on a single GPU card though.
But professional market guys! Will this finally end the dispute that the Titan wasn't made for gaming?
the Nvidia fine print always seems to have an important detail hidden away nowadays
The Titan premium has been watered down.
That immediately should cost the cost to $800, but Nv ain't having that.
I'll wait for the 980Ti.
I had high confidence that it would be SP only. NVidia learned their lesson with the DP-capable GK110 Titan - likely too many stupid high margin Quadro K5000/K6000 and Tesla K20/K40 lost sales because the pro segment was opting for almost the same capability (or more if buying the Titan Z) at a fraction of the price
I'm afraid they will be a cut down version, since Titan X is single precision, there would be no other difference.
Did you mean won't be? [Okay, you meant there will be!] I'm going to give TiX a pass to see what shakes out later.
I mean IS, since Jen Hsun just said it.
Does this mean I can't use it to mine Dogecoins?
Makes sense to introduce a cut-down Titan X as a 980 Ti. After all, NVidia has to have a $700-800 price point consumer product. The pricing scheme worked well for the 700 series, so I would not expect them to deviate too much.
970 = at around $330, it's a bargain for the performance and been selling like hotcakes since launch.
980 = at around $550, it's a little faster and cheaper than the 780Ti was.
Titan X = epeen enlargement surgery limited time only $999! Now you can prove that yours is 50% * bigger - better - faster * than the next in line product! (IRL I'm jealous towards those capable/willing to get even one )
980 Ti = suspect the specs will fall right between the 980 and Titan X (20-30% faster than 980) for around $750. Increased mem bus width of 384bit and 6GB VRAM.
980 8GB = nothing added except for more frame buffer. Probably fall somewhere in the $625-700 point. As elusive as the 6GB 780.
Take all of those $ values and subtract $100-$200 after they get #rekt by AMD in June.
I'm wetting my whistle for a sub-$400 GTX 980. Let's make it happen.
Hopefully so. Price drops are always welcome.
Shit, I should pull my 780's and get them posted FS asamfp...
Great - no transistor wasted on whole gm200 for non gaming purposes
So that means it will only cost $700, right? Going to need a source on the OP, though.
SP is fine with me, the only con might be resale value. BUT I am pretty stoked it's $999 rather than the $1,350 rumors. 12GB of VRAM ftw.
No transistors are wasted anyway...the lockout is solely an artificial software restriction. This BS is purely NVidia offering less for more. The Titan is supposed to be the be-all-to-end-all video card...now it's nothing more than a neutered mid-level card with more VRAM.
I'd even buy the damn Quadro if they wouldn't lock SLi on them. They only permit Quadro SLi only on watered down, largely useless OEM machines. The Titan was supposed to be the best of both worlds...not anymore.
NVidia's taking a play out of the Intel playbook...for non-OEM machines, give either unrestricted DP performance or SLi, but not both...sort of like Intel says you can have either an unlocked CPU or dual CPU, but not both.
Purely ignorant move on their part, IMO.
Except the new Quadro also doesn't have DP. DP was just not part of the Maxwell architecture.
I think he meant it will pay for his house in one afternoon, but I still get what he's saying.
"Huang, however, also said the card would hit roughly 200 GFLOPs in double-precision floating point performance. That puts it on par with the companys discontinuedand two generations old GeForce GTX 780 Tiin double-precision performance. Thats a break from previous Titans, which have all ruled the roost in double-precision performance. The original GeForce Titan, for example, could hit 1,500 GFLOPS.
"It is a gaming card," says analyst Jon Peddie of Jon Peddie Research. "Floating point performance is limited."
Huang seemed unconcerned with the double precision performance, however, and showed off numbers that the Titan X could perform. Used for neural net analysis, Huang said a 16-core Intel Xeon would take 43 days to process the data. The original GeForce Titan would take six days and its replacement, last years GeForce Titan Black, would take five days."
I was wondering why people were able to sell original Titan 6GB cards for $500+ after they've been out so long...now I know.
I may put mine up for sale I the base Want Ads folder to finance the purchase of a 390x in June...
Ebay is amazing for selling hardware of certain flavors. So many dumb people.
For us newbies out there- Can someone explain this like I'm five? Why is this or is this not a big deal? I'm seeing conflicting information in my research (mostly forum posts).
Are you asking for personal reasons? Because if you have to ask, it doesn't concern you.
It's only a big deal for people that want similar double precision performance and skirt around the $1600+ price tags of Quadro and Tesla offerings. For probably 99.99% of home PC gamers, it makes no difference. Think of the Titan X as providing a max power of one and a half 980's (widely varies depending on the scenario) with triple the memory.
eBay? Probably selling gt X TITAN with 12GB/2 vram cards ...
I stand corrected. The early reports made it seem that the chip's performance was artificially limited.
The one thing that gets me riled is features (or performance) artificially limited on top shelf, top price products (which is why the locked 18-core Xeons drive me so loco)...
No they took them out of GM204. Read below.
Although I understand nVidia's logic for reducing compute performance on the Titan X, for my personal use what pains me is to see that my project rendering times would increase substantially if I were to replace my eVGA GTX Titan Signature Edition with a Titan X. Sure, games would run smoother but is that a good enough reason for me to purchase a Titan X? I'll have to think about the trade-off for a bit.
That post begs the question, is it possible to use Titan X (or Titan X SLI) for gaming and have an original Titan installed as a Second/Third card dedicated to Rendering/Compute/PhysX?
If you require DP Titan X is evidently not for you. NV said it very clearly, if you need highest performance DP you buy Titan Z. Why NV did this you can see from their presentation at GPU Tech conference, they noticed there are tons of new applications requiring only SP (vision, machine learning, self-driving cars, robots). Even more, some of these applications require floating point less precise than SP and this will be supported in the Pascal architecture. NV thus relegated DP to a niche since market is very small comparing to the SP market.
You mean running simultaneously game and productivity app?
Now THIS is a fringe poster....
I was running a GTX 750 Ti as a dedicated PhysX card with my 2 x 980s in SLI (see system sig). I ended up selling it to a friend for dirt cheap as once I finished Batman AC, there was literally 0 reason to keep it anymore.
Now this guy has me thinking of adding an original Titan as my dedicated PhysX card with 2 x Titan Xs in SLI as my main (have 2 x Titan Xs arriving Wednesday to replace my 980s).
How badass would Batman AC be with that setup!!!
BTW, another reason for NV leaving DP computing might be the coming Intel Knights chip and cards. With 3 Teraflop DP performance and a version of the chip able operate as a standalone computer it might be a game changer in this area which NV can not match. If the chip would be available for making ultra high-performance workstations there would be total shakeup at the high-end (imagine Xeon on overdose of steroids) though I tend to think Intel will offer it only to exotic markets by limiting sales and/or by setting exorbitant unit price.
Blatant Intel FUD-spreading detected.
NV is not leaving dp computing. They are simply limiting it to the Tesla product line, which makes sense business-wise.
They've been pretty damn successful in scientific fields the last few years, and to claim they are pre-emptively giving it all up because of Intel vaporware is absolutely preposterous.
You know, this video card craze gets to me sometimes.
I have 980SLI and am happy with what I have, but in the back of my mind I wonder if I should get the 980TI just for the hell of it. I have no reason to get a TI model as the 980s serve me well. But this constant urge to upgrade is kinda crazy, only with GPU's. I cant be the only one, right?
I should just stop browsing the forums and keep what I have for the next 2-3 years.