Nvidia RTX 2080 Ti and RTX 2080 specifications leaked

How powerful is Turing?

Nvidia RTX 2080 Ti and RTX 2080 specifications leaked

Nvidia RTX 2080 Ti and RTX 2080 specifications leaked

When images of Nvidia's GTX 2080 Ti and GTX 2080 Ti emerged, several educated guesses could be made, especially with regards to memory bis sizes and configuration changes over Nvidia's previously announced Quadro RTX lineup

With the GTX 2080 Ti reportedly shipping with 11GB of GDDR6 memory, we could assume that the graphics card would be using a 352-bit memory bus, a cut-down version of the 384-bit bus that is present in the Quadro RTX 8000/6000. Using the same logic we could also assume that the RTX 2080 is based on the same silicon as Nvidia's Quadro RTX 5000, which employs a 256-bit memory bus, an ideal size for a GPU with 8GB of GDDR6 memory. 

Now, courtesy of Videocardz, we have detailed specifications for Nvidia's RTX 2080 Ti and RTX 2080, each of which are somewhat downgraded versions of Nvidia's Quadro RTX 6000 and 5000 respectively.

These rumoured specifications give the RTX 2080 Ti 256 fewer CUDA cores than Nvidia's Quadro RTX 6000 and 1GB less VRAM. This disparity provides Nvidia with enough wiggle room to perhaps create a new Turing-based GTX Titan GPU, which has 4,608 CUDA cores and a full 384-bit memory bus with 12GB of GDDR6 memory. 

The RTX 2080 has 128 fewer CUDA cores than Nvidia's Quadro RTX 5000, which is 384 more than Nvidia's GTX 1080 graphics card, a 15% boost in CUDA cores over its immediate predecessor. If this is combined with clock speed (or performance per clock) increases, Nvidia could perhaps yield performance improvements that are upwards of 25%, though this remains to be seen. 

 Quadro RTX 6000Geforce RTX 2080 TiGeforce GTX 1080 TiQuadro RTX 5000Geforce RTX 2080Geforce GTX 1080 
CUDA Cores4,6084,35235843,07229442560
Tensor Cores576???N/A384???N/A
GPU Memory Capacity24GB11GB11GB16GB8GB8GB
Memory Bus Size384-bit352-bit352-bit256-bit 256-bit256-bit

Nvidia is expected to reveal their Geforce RTX series of graphics cards at their "Be For the Game" Livestream on Monday, August 20th. This live stream will be available to view here

You can join the discussion on Nvidia's RTX 2080 and RTX 2080 Ti specifications on the OC3D Forums

«Prev 1 Next»

Most Recent Comments

17-08-2018, 14:18:00

Not powerful enough after over 2 years since the first launch pascal card launched.Quote

17-08-2018, 14:40:59

Originally Posted by NeverBackDown View Post
Not powerful enough after over 2 years since the first launch pascal card launched.
That depends on how well these RT cores and tensor cores pan out. We will learn more about that on monday.Quote

17-08-2018, 14:48:27

Originally Posted by NeverBackDown View Post
Not powerful enough after over 2 years since the first launch pascal card launched.
But for what? We're no longer limited by GPU horsepower for the majority of titles, so it's not that. Innovation for the sake of innovation? It's looking like Turing has a lot of that already and is by no means a refresh. A reason to upgrade? That's entirely personal. Many people didn't bother upgrading to Pascal and are on GTX 970's and 980Ti's. An RTX 2070 could offer a nice performance boost from a 970. But as for Pascal owners, if games actually demanded more performance then what it is on offer here would be more welcome as it would be more needed. Then we go back to the first question: for what? If you're on a GTX 1080 and expected huge things after two years in a market that has virtually no competition with no real die shrink, I don't know where that expectation comes from.

Quite frankly, Turing seems more than I thought it would be.Quote

17-08-2018, 16:30:50

TBH, I think the RTX 2080 Ti's reported 20+% increase in CUDA core count is a decent leap. Assuming the GDDR6 memory runs at 14Gbps, that gives a 27% boost in memory bandwidth.

Nvidia may also offer some GPU clock speed increases or a boost in performance per clock. We don't know enough to judge performance improvements right now.

Beyond that, I think that everyone is dismissing Nvidia's RT a little too quickly. Microsoft doesn't make an API for something unless they see a future in it. As an industry, we ultimately want to move towards ray tracing, Nvidia RTX won't be another PhysX.Quote

17-08-2018, 16:50:37

I sincerely doubt that they've even decided upon the Ti specs yet; I call BS. :PQuote

Register for the OC3D Newsletter

Subscribing to the OC3D newsletter will keep you up-to-date on the latest technology reviews, competitions and goings-on at Overclock3D. We won't share your email address with ANYONE, and we will only email you with updates on site news, reviews, and competitions and you can unsubscribe easily at any time.

Simply enter your name and email address into the box below and be sure to click on the links in the confirmation emails that will arrive in your e-mail shortly after to complete the registration.

If you run into any problems, just drop us a message on the forums.