Although they seemed a long way off, Nvidia's new graphics cards for 2018 could be just weeks away. Here we round up all the news and rumours surrounding the the GeForce GTX 1180 based on new Turing architecture including release date, specs and more.

Ahead of the 1180 announcement, Nvidia has confiremed Qaudro RTX cards based on Turing. We've got all the details here.

When is the GTX 1180 release date?

Not that long ago, the word from Nvidia CEO Jen-Hsun Huang at Computex 2018 cast some serious doubt over the release schedule for the latest generation of graphics cards saying that the highly anticipated GPUs won't launch for "a long time from now".

However, Nvidia has scheduled an event for 20 August at Gamescom where it is expected to announced new graphics cards. It had originally scheduled a talk entitled “NVIDIA's Next Generation Mainstream GPU” at Hot Chips on the same date.

The firm said "The event will be loaded with new, exclusive, hands-on demos of the hottest upcoming games, stage presentations from the world’s biggest games developers, and some spectacular surprises!"

The new GTX 1180, and quite possibly other cards in the new range, are expected to launch with new Turing architecture - see below.

Here are the rumoured release dates for the GTX 11 graphics cards

  • GTX 1180 – 30 August
  • GTX 1180+ – 30 September
  • GTX 1170 – 30 September
  • GTX 1160 – 30 October

If you're looking to buy new card right now, check out our roundup of the best graphics cards.

GTX 1180 or GTX 2080?

We're calling the new card the GTX 1180 following on from the other 10-series cards like the 1080, but there's a chance that Nvidia could call it the GTX 2080 instead.

Nvidia GeForce GTX 1180 specs

We've seen some rumours around the specs for the GTX 1180 circling the shadowy halls of the internet.

While we're going to be very quick to point out that these numbers are from a 'leak' and so must be taken with a very large grain of salt, they're certainly interesting to look at and could well be correct.

Card model GeForce GTX 1180 GeForce GTX 1170 GeForce GTX 1080
Architecture Turing Turing Pascal
Lithography 12nm FinFET 12nm FinFET 16nm FinFET
GPU GT104 GT104 GP104
Die Size ~400mm2 ~400mm2 314mm2
CUDA Cores 3584 2688 2560
Core Clock ~1600MHz ~1500MHz ~1607MHz
Boost Clock ~1800MHz ~1800MHz ~1733MHz
Memory 8-16 GB GDDR6 8-16 GB GDDR6 8GB GDDR5X
Memory Speed 16Gbps 16Gbps 10Gbps

Nvidia Turing

It is expected that the next GeForce GTX cards will be based on the Turing architecture - named after Alan Turing, of course. 

If you were thinking the next cards would be Volta (see below), then we did too. However, as time has gone on it seems that isn't the case. Perhaps Nvidia is wanting to keep the names separate for professional and consumer products.

Turing will be the consumer successor to Pascal, the architecture used on current 10-series GPUs.

It's now confirmed to include RTX technology - see below - and the first consumer cards with the tech are the Quadro RTX range.

"Turing features new RT Cores to accelerate ray tracing and new Tensor Cores for AI inferencing which, together for the first time, make real-time ray tracing possible." said the firm.

These RT Cores accelerate the computation of how light and sound travel in 3D environments at up to 10 GigaRays a second. Nvidia says Turing is up to 25 times faster than Pascal at real-time ray tracing operations.

It also means GPU nodes can be used for final-frame rendering for film effects at more than 30 times the speed of CPU nodes.

The Turing architecture also features Tensor Cores - processors that accelerate deep learning training and inferencing, providing up to 500 trillion tensor operations a second.

"Turing-based GPUs feature a new streaming multiprocessor (SM) architecture that adds an integer execution unit executing in parallel with the floating point datapath, and a new unified cache architecture with double the bandwidth of the previous generation."

"Combined with new graphics technologies such as variable rate shading, the Turing SM achieves unprecedented levels of performance per core. With up to 4,608 CUDA cores, Turing supports up to 16 trillion floating point operations in parallel with 16 trillion integer operations per second."

Nvidia Volta rumours - release date, specs

What about performance?

It’s way too early to say if the new generation of cards will deliver the kind of significant performance gains that Pascal did over Maxwell.

What's slightly clearer is the process the GPU will use. Intel is moving to 10nm for Cannon Lake, and that was the plan for Volta. However, shrinking the transistors is difficult (part of the reason why Intel is stuck with 14nm for four whole generations of Core processors) and the existing Volta chips ended up using a 12nm process from TSMC – the company that makes the chips for Nvidia.

The Tesla V100 and Titan V cards both use HBM - high-bandwidth memory - just like AMD's Vega cards. Previous consumer cards from Nvidia have gone with traditional GDDR RAM and while it would be nice if Nvidia kept the technically-better HBM for Volta-based cards, it may not happen.

That's because memory manufacturer SK Hynix published a press release in April 2017 which “introduced the world’s fastest 2Znm 8Gb(Gigabit) GDDR6(Graphics DDR6) DRAM. The product operates with an I/O data rate of 16Gbps(Gigabits per second) per pin, which is the industry’s fastest. With a forthcoming high-end graphics card of 384-bit I/Os, this DRAM processes up to 768GB(Gigabytes) of graphics data per second.”

It didn’t specify which manufacturer that graphics card was coming from, but with AMD’s Vega employing HBM2 memory and the press release also stating that “SK Hynix has been planning to mass produce the product for a client to release high-end graphics card by early 2018”, you’d assume it was the next-gen GeForce cards.

The Titan Xp with 12GB of GDDR5X has a memory bandwidth of 547.7GB/s, so it’s interesting that SK Hynix’s GDDR6 will run at 16Gb/s and – with a 384-bit interface card – will offer up to 768GB/s. That’s quite a jump: the Tesla V100 – with HBM2 – has a bandwidth of 900GB/s, so a putative Volta-based Titan with GDDR6 wouldn’t be too far behind.

The new Titan V - for context - offers 652.8 GB/s of bandwidth.

Nvidia Quadro RTX cards

At the Siggraph 2018 conference in Vancouver, Nvidia has revealed the first cards to be based on the Turing architecture. They come in the form of Quadro RTX models.

The firm says they will revolutionise "the work of 50 million designers and artists by enabling them to render photorealistic scenes in real time, add new AI-based capabilities to their workflows, and enjoy fluid interactivity with complex models and scenes."

There are three Quadro GPUs: the RTX 5000, RTX 6000 and RTX 8000. Nvidia also announced the RTX Server.

The new Quadro RTX cards will be released in the fourth quarter of 2018, costing around US$2,300 (around £1,800) for the RTX 5000, $6,300 (£4,940) for the RTX 6000 and $10,000 (£7,830) for the RTX 8000. 

See below for Quadro RTX specs:

GPU

Memory

Memory with NVLink

Ray Tracing

CUDA Cores

Tensor Cores

Quadro RTX 8000

48GB

96GB

10 GigaRays/sec

4,608

576

Quadro RTX 6000

24GB

48GB

10 GigaRays/sec

4,608

576

Quadro RTX 5000

16GB

32GB

6 GigaRays/sec

3,072

384

RTX real-time ray tracing

One of the coolest bits of tech that Volta supports is real-time ray tracing, a technique for rendering light that actually simulates individual rays of light as they reflect and refract off various surfaces. Ray tracing has been used in big budget Hollywood movies for years, but they have the luxury of taking hours or even days to render scenes. Games have to do it in real-time, which is what's kept ray tracing out of the gaming industry until now.

Nvidia thinks it's finally ready to change that with RTX, new ray tracing tech that - for now at least - is a Volta exclusive. First announced at GDC in March 2018, we got the chance to see it running in person at E3 2018, and came away impressed.

It seemed far off but as mentioned above, the technology is available in Nvidia's new Quadro GPUs coming later this year.

We saw two demos, both rendered in real-time by Volta GPUs, though it's worth noting that this was running off Nvidia's four-GPU, $60,000 DGX Station - which it dubs a 'personal AI supercomputer' - so it's probably safe to say your rig can't handle it yet.

The first demo we saw was built in partnership with Industrial Light and Magic, running off "experimental code" in the Unreal Engine, and showcases two First Order troopers from Star Wars on a lift ride with Captain Phasma. It really highlights the stunning reflections that RTX is capable of, showing off objects and light sources that are otherwise out of shot, and even reflecting things differently in the different types of armour.

You can watch the demo for yourself right here, but again it's worth noting that Nvidia is capable of rendering this in real-time, which is nuts. It's running at 24fps - chosen because it's the cinematic standard, helping this look as close to the actual Star Wars movies as possible:

The second demo is probably a little more indicative of how real-time ray tracing will look when it first works its way into actual games, running at 1080p and 30fps. Made in the Northlight engine by game studio Remedy (who developed  Quantum Break, and announced the upcoming Control at E3), this is another showcase for the improved reflections possible with RTX.

Since the demo was first unveiled at GDC, Nvidia has added support for a noise filter, which isn't featured in the video below, but was shown off at E3. This removes the film grain-esque noise from the below demo, making the end result much crisper and more detailed, removing some of the fuzziness you might see in the video here:

Nvidia Volta news

Nvidia Volta

After the release of the Tesla V100 and then the Titan V, speculation was justifiably rampant that a consumer version of the cards would be arriving based on the Volta architecture at some point. However, the release date of the Titan V in December has seen six months of deafening silence from camp Nvidia, which has many people scratching their heads.

Volta is the name for the microarchitecture – a design which will be used as the basis for the next generation of consumer graphics cards (as well as the workstation and enterprise stuff).

Nvidia Volta rumours - release date, specs

The current generation, including the Titan Xp and GTX 1080-1050, are based on Pascal architecture. Prior to that, it was Maxwell.

If you thought the GTX 1180 would be based on Volta then you're not alone, but it looks like it will be Turing instead.