Nvidia's 'GTX 1080 Ti' Pascal Flagship Will Be Based On The GP102 GPU - Reports Indicate 'Halfway' Point Between a GP104 and the GP100

Prototype used for illustrative purposes, not a existent Geforce poster.

Now that the initial hype of the GTX 1080 launch has died down, I thought it was time to exercise a preliminary editorial on the real Pascal flagship: the GTX 1080 Ti. We haven't published anything physical on the Pascal flagship so far and then this would serve as a baseline for future reports. The existence of the pascal chip powering this particular graphics card has been speculated upon simply nothing had come to light so far. I am talking of course about the "Gamer" version of the professional GP100 core: the GP102.

*Non a existent Nvidia logo. Feature image @Wccftech

The Nvidia GTX 1080 Ti Editorial - Powered by the GP102 GPU

Without further ado, I would like to formally admit the existence of the GP102 core. Information technology is worth pointing out that in that location have been no confirmed reports or leaks regarding this particular chip. No Zauba entries, no dice shots nevertheless. But I have heard plenty from sources familiar with these matters to say with a reasonable amount of certainty that the core exists. Non to mention that it makes a lot of sense in one case you delve into it. This will exist the core that will be present in not only the new TITAN card but the GTX 1080 Ti as well. Notice the fact however, that this mail is tagged as an editorial, so everything mentioned here is based upon my personal insights as opposed to established facts (and therefore have a non-zero probability of existence off the mark). Let us start with an introduction of the GP100 (or P100 every bit it is marketed) core.

NVIDIA GP100 Block Diagram The cake diagram of a full fledged GP100. The P100 uses a GP100 with only 56 SMs.

The GTX 1080 Ti will be powered by the GP102, which would basically be the FP32 focused cousin of the GP100. On newspaper, the GP100 had a total of 60 SMs or Streaming Multiprocessors. Each SM had a 2:1 ratio of FP32 cores to FP64 cores. This effectively meant that you were looking at 3840 CUDA Cores on the FP32 side of things and 1920 on the FP64 side. For a full of 5760 cores, the GP100 clocks in at just 610mm². Of class, as the more tech savvy would think, the P100 Accelerator didn't really use the full GP100 die. Information technology used ane with 56 SMs for a total of 3584 FP32 Cores and 1792 FP64 Cores. The reasons for the lower amount of SM tin can only exist attributed to bad yields (which are expected this early on in a node and at this large a size) which is also i more reason why we should not look the GP100 to power the GTX 1080 Ti (non a total one at any rate).

Pascal GP100 Cake diagram of a single SM from the GP100 cadre featuring a two:1 FP32 to FP64 ratio.

Of course, unlike the P100 accelerator, a gaming graphics carte du jour like the GTX 1080 Ti doesn't actually need the FP64 cores. And then there is no signal in wasting valuable dice space in hosting DP units. It is with this that nosotros tin can brainstorm the discussion of what we can wait from the GP102. Keep in mind however, that accelerators like the P100 do non take ROPs - and then we are not really looking at a linear pay-off. In that location is too the fact that the GP102 will nearly certainly not be every bit big as the GP100. Our usual sources have been very tight lipped about the this particular die simply they did country that it would be "exactly one-half fashion" between the GP104 and the GP100.

Since the GP100 and GP104 are 610mm² and 314mm² respectively, the GP102 should be in the ballpark of 462mm² to 478mm². A chip of that size should be able to host the full compliment of the GP100 FP32 cluster, or in other words 3840 FP32 CUDA cores. In any example, the amount of CUDA cores nowadays on the die will exist somewhere around this range (requite or accept a few SMs). According to our estimates, with that cadre count you are looking at power consumption within the range of 270W with GDDR5X. If Nvidia shifts to HBM2 for the GTX 1080 Ti, the consumption should fall inside the power budget of 250W. According to our sources, HBM2 will start taping out in the third quarter of this year, so it is unclear whether Nvidia will opt to become with the aforementioned.

A GP102 powered GTX 1080 Ti (or a new TITAN branded card for that affair) at the ~470mm² die size will exist able to conform to the 250W power limit and offer some pretty impressive functioning numbers. With a core count of 3840, you are looking at 50% over the raw performance of the GTX 1080 (or roughly equivalent to a GTX 1080 SLI) which would make this a solid 4K/60 capable graphics carte. I do non expect Nvidia to go for a full fledged die like the GP100 for diverse reasons (which I take detailed in my discussion below). Needless to say, the yields of a 610mm² big GPU will not exist ideal enough for mainstream purposes among other reasons. So at that place y'all go folks, the new TITAN card or the GTX 1080 Ti (or whatsoever Nvidia will call it) will exist powered past the GP102 core. As far as the when is concerned, we are even so fuzzy on the details on that count, and anything is fair game. I take touched upon some of my opinions in the complete GTX 1080 Ti discussion below, for those interested:

Nvidia GTX 1080 Ti 'Expected' Specifications Comparision

NVIDIA Graphics Card	Tesla P100	GTX 1080 Ti*	GTX 1080
Process Node	16nm	16nm	16nm
Transistors	15.3 Billion	10.viii Billion	vii.2 Billion
GPU Die Size	610 mm2	471mm²	314mm²
SMs	56	60	40
CUDA Cores Per SM	64	64	64
FP32 CUDA Cores (Total)	3584	3072-3584 (TBC)	2560
FP64 CUDA Cores / SM	32	TBD	2
FP64 CUDA Cores / GPU	1792	TBD	80
Memory Interface	4096-bit HBM2	GDDR5X	256-bit GDDR5X
Retention Size	xvi / 32 GB HBM2	x GB GDDR5X	eight GB GDDR5X
TDP	300W	250W	180W