Nvidia GeForce RTX 40 series: prices, specs, release dates and more for RTX 4080 and RTX 4090

Nvidia stuck to the high-end products in its announcement of the GeForce RTX 40 series, revealing its next-generation graphics card with only the best RTX 4090 and two RTX 4080 flavors. All three are coming soon with varying degrees of price increases over the RTX 3090 and RTX 3080 although Nvidia believes that the performance and feature enhancements of the Ada Lovelace 40 series architecture will make them worth the investment. Especially with the help of DLSS 3, a new and significantly improved version of Nvidia’s DLSS scaler that will only work with RTX 40 series GPUs at launch.

Indeed, while the next generation AMD RDNA 3 / Radeon RX 7000 series GPUs will apparently focus on improving performance, the RTX 4090 and RTX 4080 aim to increase power and AI-assisted performance gimmicks. Personally, I still hope for the less financially damaging members of the RTX 40 family – GPUs like the RTX 3060 Ti and RTX 3070 are still some of the best graphics cards you can buy – but for now, here’s all you need to know about the cards confirmed so far. .

Nvidia GeForce RTX 40 series prices

You might as well remove it quickly as the RTX 40 series’ prices are (so far) a little noteworthy:

  • RTX 4080 12 GB: £ 949 / $ 899
  • RTX 4080 16GB: £ 1269 / $ 1199
  • RTX 4090: £ 1,679 / $ 1,599

…Yes. In comparison, the RTX 3080 was launched at £ 649 / AU $ 699, while the RTX 3090 was AU $ 1,400 / AU $ 1,500, and since these are Nvidia’s starting prices, it’s likely most disc partner versions will bring you back even more. The RTX 4090 and the 16GB version of the RTX 4080 will have Nvidia-designed Founders Edition models that will sell at the prices listed above, but the “cheapest” 12GB RTX 4080 can easily break the £ 1,000 / $ 1,000 line through partner cards as well. .

On the other hand, Asus, Colorful, Gainward, Galaxy, Gigabyte, Inno3D, MSI, Palit, PNY and Zotac all produce their own RTX 40 series models. Nvidia websites.

Nvidia GeForce RTX 4090 graphics card render.

Nvidia GeForce RTX 40 series release dates

The RTX 4090 will go on sale on October 12, 2022; Hopefully for the sake of anyone looking to throw away cash straight away, this is a less grumpy launch day than the RTX 3080 when it sold out in seconds in 2020.

At the time of writing, there is no specific release date for the RTX 4080, although Nvidia says both the 12GB and 16GB variants will arrive in November. If we’re working on the RTX 4070 and / or RTX 4060, don’t expect them for a while.

Comparative image showing Microsoft Flight Simulator in native resolution and Nvidia DLSS 3.

Nvidia GeForce RTX 40 series specification

It just so happens that calling the 12GB and 16GB RTX 4080s two versions of the same GPU isn’t technically accurate: the 12GB model is based on another Ada Lovelace core chip, the Nvidia AD104, than the 16GB RTX 4080 and its AD103 processor . You can see in this specification table how both of these models are fully compared, along with the RTX 4090 based on the AD102.

RTX 4080 12 GB RTX 4080 16 GB RTX 4090 24 GB
GPU AD104 AD103 AD102
CUDA colors 7680 9728 16384
Boost the clock 2.61 GHz 2.51 GHz 2.52 GHz
Memory type GDDR6X GDDR6X GDDR6X
Memory interface 192-bit 256-bit 384-bit
Memory bandwidth 557 GB / s 742 GB / s 1018 GB / s
Energy consumption 285W 320W 450W
Power supply requirements 700W 750W 850W
Power adapter connections 2x 8-pin 3x 8-pin 3x 8-pin

Apparently there are some significant differences between the 12GB and 16GB RTX 4080 that go beyond the usual amount of memory they have in old memory. And it’s not just that prices are rising as the RTX 4090 now has the same stated power consumption and PSU power requirements as the mighty RTX 3090 Ti – though vice versa, if you already have an RTX 3080 and purchased the recommended 750W PSU, you may not need to upgrade the latter for none of the RTX 4080s.

In terms of enhancements, Ada Lovelace looks like a great architectural advance over the Ampere design of the RTX 30 series. The RT and Tensor cores (enabling ray tracing and AI / machine learning tools such as DLSS, respectively) have been redesigned to be faster and faster. more efficient. You just get more of them than you get on Ampere, and the number of CUDA cores – think of them as the staple workhorse cores for everyday play – is higher all over the world.

The RTX 40 series will not take advantage of the PCIe 5.0 slots on next-gen motherboards as Nvidia says PCIe 4.0 can still provide cards with plenty of bandwidth. That’s good – even with the upcoming release of AMD Ryzen 7000 processors and chipsets, support for PCIe 5.0 graphics on motherboards will not become the norm for a few years. RTX 4090 and RTX 4080s will also work in PCIe 3.0 slots on older mobile devices.

Collage of introductory games to support Nvidia DLSS 3.

Features of the Nvidia GeForce RTX 40 series

Faster, more numerous cores are welcome, but if the RTX 40 series proves to be worth the money, it will most likely be due to major feature enhancements.

The most interesting of all is DLSS 3. If you are not familiar yet, DLSS in its current form (the latest version is 2.4) improves gaming performance by rendering each frame at a lower resolution, thus using less power. It then applies upscaling and custom anti-aliasing – both generated by Nvidia teaching the AI ​​algorithm how to create the best images – to make each frame look like the native resolution of your monitor. It’s brilliant most of the time, and DLSS 3 goes a step further by inserting new, completely AI-generated frames between each of the “real” rendered frames.

In theory, this should yield much better FPS gains than previous DLSS versions. Each frame can also be prettier overall: the Ada Lovelace architecture includes a new “optical flow accelerator” that aims to reduce artifacts on objects when the camera is in motion. This is a rare but ugly weakness of upscalers like the DLSS and AMD FSR, and by combining this Optical Flow engine with motion vectors (data from surrounding frames that detail how the image changes between them), the DLSS 3 promises to make it upscaled games will look cleaner and cheater.

Disadvantages? One is that only a fraction of games supporting (or supporting) older versions of DLSS will also support DLSS 3. Generating new frames also increases input lag, which can be partially offset with Nvidia Reflex, but may hurt the DLSS 3’s usability in competitive scenarios . Finally, DLSS 3 is so dependent on this new optical flow accelerator that older GeForce RTX GPUs, such as the RTX 30 series and RTX 20 series, will not be able to use the new scaler when launched with new cards. Nvidia does not rule out that it will eventually release DLSS 3 for all RTX GPUs, like previous versions, though it does require a little more engineering. In short, this is the exclusive RTX 40 series for the time being.

Ray tracing has also been improved, specifically the way Ada Lovelace GPUs raise the numbers in rendering ray tracing graphics. I don’t think I can explain it without getting angry, but hang on with me.

Diagram showing how Nvidia Shader Execution Reordering on RTX 40 series GPUs.

The idea is to make the Tetris block more homogeneous.

So GPUs perform best when they perform many identical tasks in a row; say, it is faster to draw an object with multiple instances of repeating texture than if it had fewer but more varied and unique textures. Kind of like how SSDs are faster when reading data stored sequentially. A big part of why ray tracing slows down game performance so much is that it requires potential GPU performance millions various tasks, so that programmers cannot manually arrange them in a processor-friendly manner.

The RTX 40 series, however, can use a technique called Shader Execution Reordering. This allows the GPU to organize these tasks in a nice neat row, allowing itself to solve them more efficiently. Nvidia says this could improve overall performance with RT enabled by up to 25% over Ampere, which would not completely remove the FPS tax from most games, but would certainly make them run smoother.

RTX 40 GPUs will also feature AV1 encoding as standard. It won’t make your games run faster, but it’s good news if you’re interested in live streaming or making video from your game recordings: AV1 beats the current generation of formats in terms of image quality and data usage. This is bad news for Intel, however, as the AV1 encoding looked like the only true USP of their long overdue Arc Alchemist graphics cards. So much for this advantage.

Comparative image of the original Portal next to Portal with RTX, made in RTX Remix.

What is RTX Remix and what does it have to do with the RTX 40 series?

The premiere of the RTX 4090 / RTX 4080 required Nvidia to present the new RTX Remix application. Perhaps not undeservedly – if it works as intended, RTX Remix can be a very big challenge for modding older PC games.

The thing is, RTX Remix uses Ada Lovelace’s Artificial Intelligence / Machine Learning features to greatly streamline the process of creating HD mod bundles for aging games. It allows modders to import the game into the app and “capture” assets (such as textures, geometry, and lighting) with just a few clicks. These assets are then automatically converted to the USD (Universal Scene Description) file format, which allows the modder to make changes using various RTX Remix tools.

These include artificial intelligence-based texture tools that can scale up textures up to four times or detect the type of material an old texture is meant to represent – wood, metal, etc. – and automatically give them an appropriate, more detailed roughness or gloss. Ray tracing effects, Nvidia Reflex, and DLSS 3 support can also be applied, and by syncing RTX Remix with other applications such as Blender, Maya, Photoshop, and Adobe Substance, you can create completely new models and see them take shape inside the mod in real time .

Once done, RTX Remix can export everything as a mod bundle that you just need to stuff into the same folder as the game’s .exe file to run. Tons of games can potentially work with this: the only requirement, Nvidia says, is that they are either DirectX 8 or 9 based and use a fixed feature graphics pipeline.

Comparison photo of Morrowind next to his RTX Remix mod.

Here’s Morrowind Before and After Remix. The portal from RTX, seen above, was also created with Remix and runs on Steam as a free Portal DLC.

Where is the RTX 40 series coming then? For starters, you need one of these Ada Lovelace GPUs to run RTX Remix; it is not available for older Nvidia graphics cards. And while the resulting modpacks don’t have these hardware requirements, features like ray tracing and DLSS 3 clearly encourage the use (or, let’s be honest, purchase) of the RTX 4090 or RTX 4080 on the player side.

Will most modders want to make conversions that only owners of this new, overwhelmingly expensive line of graphics cards can fully benefit from? I don’t know, although anything that streamlines a complicated process, such as modding games, is likely to attract eager new users as much as it might appeal to old veterans. Nvidia is also talking to Nexus Mods and Mod DB about hosting Remix creations, so promoting and finding these mods shouldn’t be more difficult than with the handmade ones.

Leave a Reply