Tonight at a GDC party hosted by CEO Jen-Hsun Huang, NVIDIA announced the GeForce GTX 1080 Ti graphics card, coming next week for $699. Let’s dive right into the specifications!
| GTX 1080 Ti | Titan X (Pascal) | GTX 1080 | GTX 980 Ti | TITAN X | GTX 980 | R9 Fury X | R9 Fury | R9 Nano | |
|---|---|---|---|---|---|---|---|---|---|
| GPU | GP102 | GP102 | GP104 | GM200 | GM200 | GM204 | Fiji XT | Fiji Pro | Fiji XT |
| GPU Cores | 3584 | 3584 | 2560 | 2816 | 3072 | 2048 | 4096 | 3584 | 4096 |
| Base Clock | 1480 MHz | 1417 MHz | 1607 MHz | 1000 MHz | 1000 MHz | 1126 MHz | 1050 MHz | 1000 MHz | up to 1000 MHz |
| Boost Clock | 1600 MHz | 1480 MHz | 1733 MHz | 1076 MHz | 1089 MHz | 1216 MHz | – | – | – |
| Texture Units | 224 | 224 | 160 | 176 | 192 | 128 | 256 | 224 | 256 |
| ROP Units | 88 | 96 | 64 | 96 | 96 | 64 | 64 | 64 | 64 |
| Memory | 11GB | 12GB | 8GB | 6GB | 12GB | 4GB | 4GB | 4GB | 4GB |
| Memory Clock | 11000 MHz | 10000 MHz | 10000 MHz | 7000 MHz | 7000 MHz | 7000 MHz | 500 MHz | 500 MHz | 500 MHz |
| Memory Interface | 352-bit | 384-bit G5X | 256-bit G5X | 384-bit | 384-bit | 256-bit | 4096-bit (HBM) | 4096-bit (HBM) | 4096-bit (HBM) |
| Memory Bandwidth | 484 GB/s | 480 GB/s | 320 GB/s | 336 GB/s | 336 GB/s | 224 GB/s | 512 GB/s | 512 GB/s | 512 GB/s |
| TDP | 250 watts | 250 watts | 180 watts | 250 watts | 250 watts | 165 watts | 275 watts | 275 watts | 175 watts |
| Peak Compute | 10.6 TFLOPS | 10.1 TFLOPS | 8.2 TFLOPS | 5.63 TFLOPS | 6.14 TFLOPS | 4.61 TFLOPS | 8.60 TFLOPS | 7.20 TFLOPS | 8.19 TFLOPS |
| Transistor Count | 12.0B | 12.0B | 7.2B | 8.0B | 8.0B | 5.2B | 8.9B | 8.9B | 8.9B |
| Process Tech | 16nm | 16nm | 16nm | 28nm | 28nm | 28nm | 28nm | 28nm | 28nm |
| MSRP (current) | $699 | $1,200 | $599 | $649 | $999 | $499 | $649 | $549 | $499 |
The GTX 1080 Ti looks a whole lot like the TITAN X launched in August of last year. Based on the 12B transistor GP102 chip, the new GTX 1080 Ti will have 3,584 CUDA core with a 1.60 GHz Boost clock. That gives it the same processor count as Titan X but with a slightly higher clock speed which should make the new GTX 1080 Ti slightly faster by at least a few percentage points and has a 4.7% edge in base clock compute capability. It has 28 SMs, 28 geometry units, 224 texture units.
Interestingly, the memory system on the GTX 1080 Ti gets adjusted – NVIDIA has disabled a single 32-bit memory controller to give the card a total of 352-bit wide bus and an odd-sounding 11GB memory capacity. The ROP count also drops to 88 units. Speaking of 11, the memory clock on the G5X implementation on GTX 1080 Ti will now run at 11 Gbps, a boost available to NVIDIA thanks to a chip revision from Micron and improvements to equalization and reverse signal distortion.
The TDP of the new part is 250 watts, falling between the Titan X and the GTX 1080. That’s an interesting move considering that the GP102 was running at 250 watts with identical to the Titan product. The cooler has been improved compared to the GTX 1080, offering quieter fan speeds and lower temperatures when operating at the same power envelope.
Performance estimates from NVIDIA put the GTX 1080 Ti about 35% faster than the GTX 1080, the largest “kicker performance increase” that we have seen from a flagship Ti launch.
Pricing is going to be set at $699 so don't expect to find this in any budget builds. But for the top performing GeForce card on the market, it's what we expect. It should be on virtual shelves starting next week.
(Side note, with the GTX 1080 getting a $100 price drop tonight, I think we'll find this new lineup very compelling to enthusiasts.)
NVIDIA did finally detail its tiled caching rendering technique. We'll be diving more into that in a separate article with a little more time for research.
One more thing…
In another interesting move, NVIDIA is going to be offering “overclocked” versions of the GTX 1080 and GTX 1060 with +1 Gbps memory speeds. Partners will be offering them with some undisclosed price premium.
I don’t know how much performance this will give us but it’s clear that NVIDIA is preparing its lineup for the upcoming AMD Vega release.
We’ll have more news from NVIDIA and GDC as it comes!
I’ll take two.
I’ll take two.
Fan of last year technology?
Fan of last year technology?
Nope its the same Pascal
Nope its the same Pascal architecture as the GTX-1080, nVidia got the GP102 a lot earlier out then the GP104.
I am getting 2 of these for my VR, upgrading from my now aging Quad SLI Titan v1 setup.
Interesting. whit the faster memory the 980Ti should also have a 10% better memory latency then the Titan X (Pascal).
Irritating that you have two parts called Titan X, whit Pascal and Maxwell SKU’s! >_<
And you just know they were
And you just know they were pissed that AMD didn’t reveal much more about Vega today. Which, to be honest, kind of puts the ball back in Raja’s court and that’s an enviable position to be in, although Nvidia is going to sell a heck of a lot of 1080 and 1080 Ti hardware between now and then.
Yea… that price point is
Yea… that price point is going to pickup a lot of impulse buys. Especially for those who were already in the market for a new chip, and were merely waiting for these announcements.
And yes, from a certain point, amd is in the enviable position. But what does this say about Vegas performance, if anything?
It doesn’t say anything about
It doesn’t say anything about it other than, if they are targeting 1080 performance, they have to come in under $400 to be competitive in 2H 2017. If they are targeting 1080 Ti performance, they can come in around $500 and the price war will start.
Worst case scenario for AMD is they come out with something that beats 1080 Ti (I think all signs point to this being extremely unlikely, but it would be nice to see) and Volta drops soon after that, shattering any lead they may have.
If that does happen, though, we’d be back in the leap frog scenario we used to see, which was good for competition and the respective egos of AMD and Nvidia fans everywhere.
nvidia can probably just use
nvidia can probably just use fully enabled GP102 with very high stock clock. imagine 1.8ghz stock with initial boost to 2.0ghz. overclock headroom will be limited or none existent at all but most people out there only look at “reference performance”. even most reviewer still using reference card despite majority people out there buying factory overclocked card. if that still cannot beat vega then nvidia will have to release bigger pascal. remember GP102 is still considered as small at 471mm2 unlike GP100 which measured at 610mm2.
The GP100 is build pure fore
The GP100 is build pure fore compute work, yes it will properly work also well for gaming, but its properly not worth the hassle for nVidia to make a 1090Ti or Titan X+.
As the yields are much lower, and the architecture is not build for gaming, i believe it has not bin optimized for VR or gaming.
Just like Intel Core 7700K is in most cases quicker then a i7-6x00K, the GP100 will properly also have gaming efficient problems, that would properly make it quicker then a GP100, but less then the transistor count would suggest.
Lol @ the Titan XP owners!
Lol @ the Titan XP owners!
Why? Early adopter tax, and
Why? Early adopter tax, and they’ve had many, many months with their hardware. Now if someone just bought a Titan X yesterday, they should be looking to return it…
TheWayYou’reMeantToBeDripFedI
TheWayYou’reMeantToBeDripFedInaneDetailsForAYearAndAHalf
at least they had a physical product for which to shell out
but yea, this new price point is rough
Titan owners have been
Titan owners have been enjoying ti performance for almost year now
So not only did con customers
So not only did con customers with crippled Titan X(pascal) but now they are rubbing salt to injury by selling that same chip at lower price.
why people still going with
why people still going with this? those that buying titan(s) know what they get into. cheaper alternative at the same (or even better performance) will always trail titan release few months later. it has been that way since 2013.
When will AIBs can start
When will AIBs can start selling? A month later? Is this still called Founder’s Edition?
I think I was the only one
I think I was the only one predicting the 1080ti will be similar to the Titan x. To make enthusiast and gamers believe they are getting a bargain meanwhile this was Nvidia’s plan the past 2 generations.
No, you weren’t, not by a
No, you weren’t, not by a mile. Everyone with a brain saw this coming. Intel does it, Apple does it, Tesla does it, Samsung does it, Yamaha does it – everyone tries to do it! It’s basic economies of scale.
Release an expensive, high end product with decent margin while yields are low and demand is high. Build up cash and refine the process. When demand slows down and margins and yields are higher, release the refined product for less money to drive up demand and sell volume.
This is the third generation of nvidia using the same exact playbook. Why does no one understand that this is just good business?
Most tech sites said it was
Most tech sites said it was going to be cut down chip lol with performance sitting between 1080 and Titan x. I’m not saying it’s not clever for the business aspect. As a consumer I care less of the profits. Only apple and Nvidia trolls care about profits and if you are a shareholder. Some neive consumers believe that the pricing is actually a bargain meanwhile it’s actually the monopolied price Nvidia set.
No HBM2 ? Fail. Ngredia is
No HBM2 ? Fail. Ngredia is going to get recked by Vega as hard as Intel is by Ryzen.
k
k
No DVI port but they will
No DVI port but they will include an adapter? Will this be a dual link DVI adapter? Will it work for older Nvidia lightboost displays?
Don’t know but I like the
Don’t know but I like the change. Maybe we can finnaly get a single slot backplate for watercooling.
I’ll expect to see some
I’ll expect to see some competition between AMD and NVidia.
If they had to use the performance of a titan X for the ti and then price it at $699, you can assume than Vega top of the line will be slightly slower than the titan X /1080ti (5%) for $50 to $80 less. (And do better with like 5% to 10% with Vulcan or DX12 games)
Which is not that impressive for a long awaited card…
This is simply amazing. AMD
This is simply amazing. AMD outdone itself. They basically forced noVideo/nGreedia to spill out all of their beans, while AMD itself only shown /Ve/Ga/’s performance improvements, new technologies, and the actual name of the card. noVideo couldn’t handle the pressure and not only lowered 1080’s price, but also eagerly jumped out with pricing of 1080Ti. All AMD needs to do now to completely obliterate noVideo, is just release /Ve/Ga/ at ~400$, because /Ve/Ga/ already performs better than 1080. And now literally nothing stops AMD from doing it. Just that one thing and nothing else. And AMD has all the time on their hands, due to the release being slated for summer. Bravo, AMD. Bravo. I applaud while standing. We all do. What a truly great time to be alive as a hardcore PC enthusiast in general and a PC user in overall. Glorious. Astounding. Excellent. Magnificent. Fabulous. Awesome. Marvelous. Fantastic. Perfect.
Are you on glue?
Are you on glue?
Cry harder, noVideot.
Cry harder, noVideot.
You’re hilarious dude
You’re hilarious dude
I may be “hilarious”, but at
I may be “hilarious”, but at least I’m not a brainless zombified noVideotic milking sheep.
http://www.thesmokinggun.com/
http://www.thesmokinggun.com/sites/default/files/imagecache/750×970/documents/0721051gold1.jpg
I think this is his mugshot
You should be spanked for the
You should be spanked for the kind of asshattering idiocy you’re producing right now. Grow up already.
So the guy who posted “Are
So the guy who posted “Are you on glue?” gets a free pass? why don’t you go spank him too?
https://healingfromcomplextra
https://healingfromcomplextraumaandptsd.files.wordpress.com/2016/01/spanking-is-abuse.png?w=611
So you’re one of them
So you’re one of them “juvenile system”-brainwashed liberals, I see. My deepest condolences to your relatives. You’re beyond repair.
Where did I say that he does?
Where did I say that he does?
your lack of response to the
your lack of response to the first guy.
You are just as much of an asshat as the guy you called an asshat!
Except RTG does not have a
Except RTG does not have a better or even equal card to the 1080 Ti.
If you remember correct when the showed up the 4K Ryzen on BF1 they used Titan X(P) cause ….cause there’s nothing else that can drive it 🙂
Yeah but what will the profit
Yeah but what will the profit margins be on Vega at $400 with expensive HBM2? This is a repeat of Fury launch all over again. That’s what AMD gets for being late to the dinner table again. Scraps and leftovers.
Maybe AMD will scam enough people into buying with their canned benchmarks again.
Go away Chizow.
Go away Chizow.
I was lucky and got a 1080
I was lucky and got a 1080 last year for 530 so im not salty about that price drop. definitely the best card I have ever owned.
This reads soooooo much like
This reads soooooo much like an Apple type move to squeeze the most money from their customers who blindly follow along. isheep & nsheep?
Why not both?
insheep
Of
Why not both?
insheep
Of wait… That is what both companies are doing…
11GB?
8 good + 3 crap?
11GB?
8 good + 3 crap?
I might be mistaken on this,
I might be mistaken on this, but I believe it has to do with the number of ROPs. Take a look at the titan xp, it has 96 ROPs and 12Gb of vram, which equals to 1GB of ram per 8 ROPs. so, 1080TI has 88 ROPS, so 88/8 = 11, so half life 3 confirmed.
11GB seems soooo weird, I had
11GB seems soooo weird, I had to read it twice, especially mixed with the 11Ghz spec, thought it was a typo.
More like 10.5
More like 10.5
I(AMDeficient) fanboy
I(AMDeficient) fanboy
You’re not trying hard
You’re not trying hard enough, kiddo.
Wonder how long until someone
Wonder how long until someone figures out how to unlock those “disabled” units.
Assuming these new tweaks on
Assuming these new tweaks on 1080 will be able to achieve with current 1080 with vbios flash
Looks like Intel & Nvidia can
Looks like Intel & Nvidia can just stop their price gouging and all of AMD’s competitive advantage is gone. AMD might have a better solution than Intel but with this slow release of their competitive $200 parts is it going to be to late? That’s like 6 months away, no?
As for Nvidia, Nvidia is going to have a new GPU too.
11.000 MHz memory clock!!!!
11.000 MHz memory clock!!!! LoL!!!
Stop fooling your readers with those numbers Ryan.
11GB…..Anytime Nvidia does
11GB…..Anytime Nvidia does VRAM tricks I avoid even caring.
No tricks actually, the
No tricks actually, the memory bus is smaller than the titan X, and therefor it has exactly the amount of memory it should:
Titan X: 384 / 12 = 32
1080 Ti: 352 / 11 = 32
Same goes for the ROP units:
Titan X: 96 / 12 = 8
1080 Ti: 88 / 11 = 8
So 32bits, and 8 ROP units per GB of ram, no tricks being used here at all.
Nvidia’s basically saying a
Nvidia’s basically saying a few things
1. 1070 gimped heres a few bucks off
2. 1080 was overpriced, heres $200 off
3. 1080 gen2 = 680 ~ 770 throwback…… throw in more VRAM 12/16GB, could be like the old 770 4GB cards I still have