GM200 Specifications
NVIDIA has let loose the GM200 GPU upon the world, in the form of the GTX TITAN X. It has 12GB of memroy…12GB!!
With the release of the GeForce GTX 980 back in September of 2014, NVIDIA took the lead in performance with single GPU graphics cards. The GTX 980 and GTX 970 were both impressive options. The GTX 970 offered better performance than the R9 290 as did the GTX 980 compared to the R9 290X; on top of that, both did so while running at lower power consumption and while including new features like DX12 feature level support, HDMI 2.0 and MFAA (multi-frame antialiasing). Because of those factors, the GTX 980 and GTX 970 were fantastic sellers, helping to push NVIDIA’s market share over 75% as of the 4th quarter of 2014.
But in the back of our mind, and in the minds of many NVIDIA fans, we knew that the company had another GPU it was holding on to: the bigger, badder version of Maxwell. The only question was going to be WHEN the company would release it and sell us a new flagship GeForce card. In most instances, this decision is based on the competitive landscape, such as when AMD might be finally updating its Radeon R9 290X Hawaii family of products with the rumored R9 390X. Perhaps NVIDIA is tired of waiting or maybe the strategy is to launch soon before Fiji GPUs make their debut. Either way, NVIDIA officially took the wraps off of the new GeForce GTX TITAN X at the Game Developers Conference two weeks ago.
At the session hosted by Epic Games’ Tim Sweeney, NVIDIA CEO Jen-Hsun Huang arrived when Tim lamented about needing more GPU horsepower for their UE4 content. In his hands he had the first TITAN X GPU and talked about only a couple of specifications: the card would have 12GB of memory and it would be based on a GPU with 8 billion transistors.
Since that day, you have likely seen picture after picture, rumor after rumor, about specifications, pricing and performance. Wait no longer: the GeForce GTX TITAN X is here. With a $999 price tag and a GPU with 3072 CUDA cores, we clearly have a new king of the court.
GM200 GPU Specifications
The basis for the new GeForce GTX TITAN X is NVIDIA’s GM200 GPU. This large, beastly part is based on the same Maxwell architecture that we know and love from the GTX 980, GTX 970 and GTX 960. It is also built on the same 28nm process technology that those same GPUs use, too. NVIDIA is not yet moving over to another process tech quite yet, despite the rumors that AMD is going to be migrating to 20nm on its next flagship GPU.
The GM200 includes 3072 CUDA cores, 192 texture units and a 384-bit memory bus. Clearly this GPU isn’t bluffing, it has some power.
TITAN X | GTX 980 | TITAN Black | R9 290X | |
---|---|---|---|---|
GPU | GM200 | GM204 | GK110 | Hawaii XT |
GPU Cores | 3072 | 2048 | 2880 | 2816 |
Rated Clock | 1000 MHz | 1126 MHz | 889 MHz | 1000 MHz |
Texture Units | 192 | 128 | 240 | 176 |
ROP Units | 96 | 64 | 48 | 64 |
Memory | 12GB | 4GB | 6GB | 4GB |
Memory Clock | 7000 MHz | 7000 MHz | 7000 MHz | 5000 MHz |
Memory Interface | 384-bit | 256-bit | 384-bit | 512-bit |
Memory Bandwidth | 336 GB/s | 224 GB/s | 336 GB/s | 320 GB/s |
TDP | 250 watts | 165 watts | 250 watts | 290 watts |
Peak Compute | 6.14 TFLOPS | 4.61 TFLOPS | 5.1 TFLOPS | 5.63 TFLOPS |
Transistor Count | 8.0B | 5.2B | 7.1B | 6.2B |
Process Tech | 28nm | 28nm | 28nm | 28nm |
MSRP (current) | $999 | $549 | $999 | $359 |
Essentially, the GTX TITAN X’s compute structure is exactly a 50% boost over the GeForce GTX 980 with just a slight reduction in clock rates at stock settings. You get 50% more processing cores, 50% more texture units, 50% more ROP units, 50% larger L2 cache and a 50% larger memory bus. Dang – that’s going to provide some impressive computing power, resulting in a peak theoretical throughput of 6.14 TFLOPS single precision.
During the keynote at GTC, NVIDIA's CEO quoted the single precision compute performance as 7.0 TFLOPS, which differs from the table above. The 6.14 TFLOPS rating above is based on the base clock of the GPU while the 7.0 TFLOPS number is based on "peak" clock rate. Also, just for reference, at the rated Boost clock the Titan X is rated at 6.60 TFLOPS.
A unique characteristic of this TITAN X card is that it does not have an accelerated performance configuration for double precision computing, which is something that both the TITAN and the TITAN Black had before it. The double precision performance is still a 1/32nd ratio (relative to single precision). That gives the TITAN X DP compute capability at just 192 GFLOPS. For reference, the TITAN Black has DP performance rated at 1707 GFLOPS with a 1/3rd ratio of the GPU’s 5.12 TFLOPS single precision capability. It appears that NVIDIA is not simply disabling the double precision compute capability on the GM200 GPU, hiding it and saving it for another implementation. Based on the die size, shader count and transistor count, it looks GM200 just doesn't have it. NVIDIA must have another solution up it's sleeve for Maxwell double precision compute.
Oh, and yes, before you ask me in the comments below, I have directly asked NVIDIA for comment about memory configuration concerns, in regards to the GTX 970. The TITAN X does not have any divided memory pools, the 384-bit memory bus is not sectioned off into any sub-groups that may or may not run at slower throughput. So, there’s that.
The GM200 is built on 24 SMX modules and this is the full GPU implementation so you should not expect another variant to show up down the road with anything higher than 3072 CUDA cores, unless the company spins a completely new GPU revision. Being that the GPU is still based on the 28nm TSMC process, and is compiled with 8 billion transistors, this is not a small component. Measured die size was 25mm x 25mm or 625mm2. The previous largest GPU we had seen is the GK110 used 7.1 billion transistors and had a die size of around 561mm2. (Note that NVIDIA sets the die size at 601mm2 based on measurements of 24.66mm x 24.38mm.)
Clock speeds on the TITAN X are lower than the GTX 980 at stock, as you would expect, but the differences aren’t drastic. With a base clock of 1000 MHz and rated Boost clock of 1075 MHz. The GTX 980 reference clocks are 1126 MHz/1216 MHz respectively resulting in a 12.6% decrease comparatively. The memory clock is still starting at 7.0 GHz, the same speed as the GTX 980 and even the previous GTX TITAN Black. The memory configuration is improved with 50% higher peak bandwidth, hitting 336.5 GB/s, which is higher than the 320 GB/s rated by the AMD Radeon R9 290X flagship.
Speaking of memory, the TITAN X will ship with 12GB of memory. Gulp. That is 3x the memory found on the GTX 980 and 2x the GTX TITAN Black which seemed crazy (at 6GB) when it shipped in February of 2014. I can already hear the debate and no, 12GB of memory is not going to be beneficial for gamers for quite some time – likely a time span even further out than the life of this GPU, to be fair. With the recently debate around the 3.5GB and 4GB frame buffers, we saw that most games available today, even when pushed with settings intending to increase memory usage, rarely extend in the world of 4GB. If you want to think really crazy, let’s assume you are planning on doing on 4K Surround gaming with three displays; you might be able to stretch that out to 7-8GB if you try real hard. Of course there is a secondary audience for this card that focus on GPGPU compute rather than gaming, where that 12GB of memory could be more useful, more quickly.
The only negative change on the GM200 (compared to GM204) is its rated TDP. With a listing of 250 watts, the GTX TITAN X is clearly going to run hotter than the GTX 980, which has a 165 watt TDP. (Of note, this is a 51% increase in TDP that is in line with the other specification changes.) Keep in mind that the Radeon R9 290X has a TDP of 290 watts, though we have measured it higher than that in our power testing several times. Both the new TITAN X and the R9 290X require a 6+8 pin power connection so it will be interesting to see how the real-world power consumption varies between these two GPUs.
The new GTX TITAN X shares the same feature set and capability as the GTX 980. That includes support for MFAA, VXGI acceleration, Dynamic Super Resolution, VR Direct and more. For more information on those features and what they bring to the table check out the links below.
Now, let’s dive into the design of the GTX TITAN X graphics card itself!
very overpriced for what you
very overpriced for what you get, I was hoping they’d incorporate a 512bit memory interface and stick with double precision
They’re already pushing over
They’re already pushing over 600mm2 on the GPU. They seem limited by what they can do on the 28nm process.
Oh…oh god…MOTHER OF
Oh…oh god…MOTHER OF GOD…
http://www.imagebam.com/image/4dccee397912995.
It seems like most newer
It seems like most newer games work quite well with multiple gpu set-ups. At some point there is going to be little difference between 2 gpus and one gpu at twice the size except price. Two smaller gpus will probably be cheaper since the yield on a ~600 square mm part is not going to be good. For most of the games tested, a multi-gpu system is cheaper and better performing. Power is more, but with multi-gpu systems you are running multiple, independent high speed memory systems instead of a single memory system. Stacked HBM is supposed to be much more power efficient, so multiple smaller gpus with HBM may be the the best option eventually. Both the Titan X and the 390X are going to be priced out of the mainstream market. They are interesting to read about, but I doubt many people here are considering actually buying one.
Performance per dollar is
Performance per dollar is HORRIBLE, this is the card for people who have so much money they honestly don’t know what to do with it.
“As a hardware enthusiast,
“As a hardware enthusiast, it’s impossible to not fall in love with the GeForce GTX Titan X.”
Yes it is. Because despite what your “funny nvidia infomercial”, which you dare call a test, this card is a freaking rip-off.
It’s nothing more than this generation’s GTX580 3GB. Except that one was 550$. It’s gain over previous gen, be it performance or power efficiency, are nothing more than 780 over 580 or 580 over 280.
The fact the GTX 560Ti was 250€ and the 580 3GB 550€, and now the 960 Ti (oop, 980) is 550€ and the 980 (oops, TitanX) 1000€ should make any hardware enthusiast think twice before writing piles of advertising shit.
Wow, angry much ?
AMD
Wow, angry much ?
AMD overclocked one of their stupid bull cpu’s, then sold it on NewEgg for freaking $800.
I was LOL’ing so hard – a bunch of amd fanboys bought it, then it found to be flaky and a supreme power hog, then the price plummeted… to like $300, slightly above the best doggy cpu ad had
I mean how could amd rebrand then raise the price 300% ?
Now THERE’S A RIPOFF ! IT WASN’T EVEN A NEW CPU !
At least nVidia makes a new MONSTER product with MONSTER RAM, instead of overclocking a failed dog then scalping fanboys.
The 295X2 just killed another
The 295X2 just killed another weak Titan.
The 295X2 is still top of the food chain.
The weak Titan X is too weak for 295X2.
AMD continues to bleed market
AMD continues to bleed market share and profits on discrete GPUs.
Nvidia gains market share in discrete and now holds a 76% to 24% advantage over AMD.
http://jonpeddie.com/news/comments/gpu-shipments-marketwatch-q4-2014-charts-and-images
AMD cuts the prices over and over again yet more gamers are choosing Nvidia over AMD.
Nvidia has high margins on the Titans (and they do sell out) whereas AMD has low margins on the 295X2 (2x GPUs, expensive cooling, more layers on PCB, etc) and still can’t get people to buy it. So sad (for AMD).
Fanboys make me laugh. If you
Fanboys make me laugh. If you don’t want what the Titan X offers, don’t buy it. If you can’t afford it, don’t buy it. But laughing and saying the 295X2 is still the best is just stupid. The 295X2 is 12.4B transistors and 600W of power that can only keep itself cool with an unwieldy CLC. AMD’s only selling for 700 bucks because nobody wants them. Is it still the fastest single PCB solution? Well, yes, but only sometimes… only if CF is implemented correctly. The 295X2 comes with so many compromises that nvidia clearly feels like they can charge a $450 premium over the GTX 980, and until AMD has something that can compete directly on a single GPU basis, its going to stay that way. Oh, and a bunch of sketchy leaked slides and rumors about the 390X mean nothing. If you’re using the 390X as proof of AMD’s superiority you’re a joke. Call me when the reviews drop.
PS- owner of 2 290x’s speaking
Prices at this level have
Prices at this level have very little to do with market demand or much of anything else. The volume of these super ridiculous high end cards is so low that they are mostly a marketing tool. Nvidia releases this Titan x with probably 2x the amount of ram that it could actually use and gets a lot of media attention ahead of any AMD releases. They are not going to make much on actually selling these things, especially since it doesn’t have high compute performance. They will certainly have a cut down version of this chip that will be a more reasonable price eventually.
For people actually willing to spend the money on this much performance, the r9-295×2 is still the better deal. I don’t care much about the number of transistors and such. The AMD solution is cheaper and performs better. Most gamers don’t care that much about the power consumption either unless it makes the card excessively noisy. I don’t see much of any reason that a gamer would buy a Titan x over the 295, so why wouldn’t you say that the 295 is still the better card? It seems that most modern games are handling dual gpu fine, and older games probably don’t need this level of performance anyway.
Agreed, despite your being
Agreed, despite your being too cowardly to register.
The transistor counts tell
The transistor counts tell you something about where each company is in terms of technology, but you’re right that they don’t matter directly. Instead it has a knock-on effect for the power and cooling requirements. Despite what AMD fanboys keep telling me, power DOES matter. Installing a CLC is a big logistical pain in the ass and fan speed needs to compensate to keep things cool. I couldn’t stand the fan noise on my 7970’s, let alone the 290X’s so I built a full custom loop just to shut them the hell up. Might not have needed to do that with more efficient cards that can run quieter on air.
And as far as performance, I have lots of recent games that still only use 1 GPU reliably. How many games have serious, sometimes gamebreaking issues at launch with SLI or CF? Happens a lot… flickering textures, stuttering, framerates bouncing all over the place or just straight up crashes. Titanfall and Evolve come to mind immediately. So yea, if everything is working the 295X2 is a better perf/$ card, but its so much more dependent on software that there will always be value in a single GPU solution. How much is up to you, but people who keep squawking about the 295X2 would have you believe otherwise.
So now crossfire always works
So now crossfire always works and amd drivers don’t suck….
SELL IT TO THE CARPETBAGGER BRO.
Stop talking sense!
Stop talking sense!
deleted. Was going to say
deleted. Was going to say deleted self, but that would have been ambiguous.
Yes the 295x has compromises
Yes the 295x has compromises
BUT I want one, but it won’t fit into the cases I own, as I already have watercooling CPU units inside. If they were Air cooled they would be flying off the shelf at that price point. Thats why most AMD ers are just doing crossfire.
PS. (I run AMD and Nvidia)
Priced here in Ireland @ 1250
Priced here in Ireland @ 1250 euros (Including tax 23%).
I want one, (Well 2 in SLI), but not ever at that price.
980 TI is basically going to be this card with less ram and speed aligned as to what the AMD 390x does. Prob june/july
So wait and see, seems to be the thing to do.
(If rumors are true then the TITAN X is about AMD 390x speeds)
PS. Never seen so may review cards issued………..
PPS. Guess I didn’t win the draw………..secret word was dress (Red)
I see a future with 3-phase
I see a future with 3-phase 208 and liquid nitrogen tanks.
Comon guys, Titan X
Comon guys, Titan X benchmarks on vanilla Skyrim? 🙁 Nobody buying Titan X is going to care about vanilla Skyrim performance on 1440p. Maybe 4k vanilla, but better, 1440p with mods.
Want to post a score for EVGA
Want to post a score for EVGA super-clocked version of this card. With factory settings and without additional overclocking I got on Fire Strike Extreme: Score: 8281, Graphics score: 8884 which is about 10% higher that reference card.