Sticking with the high-end products for the announcement of their GeForce RTX 40 series, Nvidia unveiled their next-generation graphics cards featuring only the premium RTX 4090 and two flavors of RTX 4080. All three are coming soon with varying degrees of price increase over the years. of the RTX 3090 and RTX 3080, although Nvidia believes the performance and feature improvements of the 40-series Ada Lovelace architecture will make them a worthwhile investment. Mainly using DLSS 3, a new and significantly improved version of Nvidia’s DLSS upscaler that will only work with RTX 40 series GPUs at launch.
Indeed, while AMD’s next-gen RDNA 3/Radeon RX 7000 series GPUs will apparently focus on efficiency gains, the RTX 4090 and RTX 4080 are all about going big in power and AI-assisted performance tricks. Personally, I’m still hoping for less financially devastating members of the RTX 40 family — GPUs like the RTX 3060 Ti and RTX 3070 are still some of the best graphics cards money can buy — but for now, here’s everything you need to know about the confirmed cards so far.
Nvidia GeForce RTX 40 Series Pricing
Might as well get this out of the way quickly, as the price of the RTX 40 series (so far) is quite hair-raising:
RTX 4080 12GB: £949 / $899 RTX 4080 16GB: £1269 / $1199 RTX 4090: £1679 / $1599
…Yes. By comparison, the RTX 3080 launched for £649/$699, while the RTX 3090 was $1400/$1500, and since these are Nvidia’s starting prices, it’s likely most board partner versions will hold you back even more. The RTX 4090 and the 16GB version of the RTX 4080 will have Nvidia-designed Founders Edition models, which will sell at the prices listed above, but the ‘cheapest’ 12GB RTX 4080 could easily top the £1000/$1000 line. can break through partner cards as well.
Speaking of Asus, Colorful, Gainward, Galaxy, Gigabyte, Inno3D, MSI, Palit, PNY and Zotac, they all produce their own RTX 40 series models. Not EVGA, mind, abandoning the graphics card business entirely, citing uncooperative practices on Nvidia’s part.
Nvidia GeForce RTX 40 Series Release Dates
The RTX 4090 goes on sale October 12, 2022; I hope, for the sake of anyone wanting to drop the cash right away, it’s less of a teeth-gritting launch day than the RTX 3080 had when it sold out in seconds in 2020.
There is no specific release date for the RTX 4080 at the time of writing, although Nvidia says both the 12GB and 16GB variants will launch in November. If an RTX 4070 and/or RTX 4060 is in the works, don’t expect them for now.
Nvidia GeForce RTX 40 series specifications
Coincidentally, it’s technically incorrect to call the 12GB and 16GB RTX 4080s two versions of the same GPU: the 12GB model is based on a different underlying Ada Lovelace chip, the Nvidia AD104, than the 16GB RTX 4080 and its AD103 processor. . You can see in this specs table how the two compare in full, along with the AD102-based RTX 4090.
RTX 4080 12GB RTX 4080 16GB RTX 4090 24GB GPU AD104 AD103 AD102 CUDA cores 7680 9728 16384 Boost clock 2.61GHz 2.51GHz 2.52GHz Memory type GDDR6X GDDR6X GDDR6X Memory interface 192-bit 256-bit 384-bit Memory bandwidth 557GB/s 742GB/s 1018GB/s Power Consumption 285W 320W 450W Power Requirements 700W 750W 850W PSU Connectors 2x 8-pin 3x 8-pin 3x 8-pin
Apparently there are some significant differences between the 12GB and 16GB RTX 4080s beyond just how much memory they have in the old noggin. And it’s not just prices that are creeping up, as the RTX 4090 now has the same quoted power consumption and PSU wattage requirement as the mighty RTX 3090 Ti – although vice versa, if you already own an RTX 3080 and bought the recommended 750W PSU, you don’t need to upgrade the latter for any of the RTX 4080s.
As far as upgrades go, Ada Lovelace looks like a major architectural advancement over the RTX 30 series Ampere design. The RT and Tensor cores (which respectively enable ray tracing and AI/machine learning tools such as DLSS) have been redesigned to be faster and more efficient. You also just get more of it than on Ampere, and the number of CUDA cores — think of these as the basic workhorse cores for everyday gaming — is higher across the board.
The RTX 40 series won’t use PCIe 5.0 slots on next-gen motherboards, as Nvidia says PCIe 4.0 can still provide the cards with plenty of bandwidth. That’s fine – even with the imminent launch of the AMD Ryzen 7000 CPUs and chipsets, PCIe 5.0 graphics support on motherboards is several years away from the norm. The RTX 4090 and RTX 4080s also work in the PCIe 3.0 slots of older mobos.
Nvidia GeForce RTX 40 Series Features
Faster, more cores are welcome, but if the RTX 40 series eventually becomes worth the money, it will likely be thanks to the major feature upgrades.
Most interesting of all is DLSS 3. If you’re not already familiar with it, DLSS in its current form (the latest version is 2.4) improves performance in games by rendering each frame at a lower resolution, consuming less horsepower . It then applies upscaling and custom anti-aliasing — both generated by Nvidia teaching an AI algorithm how to produce the best images — to make each frame look like your monitor’s native resolution. It’s mostly brilliant, and DLSS 3 goes a big step further by inserting new, fully AI-generated frames between each of the ‘real’ rendered frames.
In theory, this should deliver significantly better frames per second than previous DLSS versions. Every frame could also be more beautiful 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 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), DLSS 3 promises to make upscaled games look cleaner and better. to show off. sharper.
cons? One is that only a fraction of the games that support (or will support) older DLSS versions will also support DLSS 3. Generating new frames also increases input lag, which can be partially offset with Nvidia Reflex, but could hurt the usefulness of DLSS 3 in competitive scenarios. Finally, DLSS 3 is so dependent on this new Optical Flow Accelerator that older GeForce RTX GPUs i.e. RTX 30 series and RTX 20 series will no longer be able to use the new upscaler once it launches alongside the new cards. Nvidia doesn’t rule out DLSS 3 eventually becoming available for all RTX GPUs, as previous versions are, although it will require a bit more engineering. In short, it is an exclusive RTX 40 series for now.
Ray tracing is also getting an upgrade, particularly in how Ada Lovelace GPUs are cracking the numbers when rendering ray-tracing graphics. I don’t think I can explain this without getting a little goofy, but be patient.
It’s about making the Tetris block more homogeneous.
So graphics processors work best when they perform many identical tasks in succession; it is faster to draw, for example, an object with many instances of a repeating texture than if it had fewer, but more varied and unique textures. Kind of like how SSDs are faster at reading data stored sequentially. A big part of the reason ray tracing slows down game performance so much is that it requires the GPU to perform potentially millions of different tasks, enough that developers can’t manually order them in a processor-friendly way.
However, the RTX 40 series can use a technique called Shader Execution Reordering. This allows the GPU itself to arrange these tasks in a nice, neat row, allowing it to tackle them more efficiently on its own. Nvidia says this can improve overall performance with RT enabled by up to 25% over Ampere, which wouldn’t completely wipe out the FPS load in most games, but would certainly make them run smoother.
RTX 40 GPUs also have standard AV1 encoding capabilities. This won’t make your games run faster, but it’s good news if you’re into live streaming or shooting video from game footage: the AV1 format beats the pants of current-gen formats for image quality and data usage. It’s bad news for Intel, though, as AV1 encoding looked like the one and only USP of their highly anticipated Arc Alchemist graphics cards. So much for that benefit.
What is RTX Remix and what does it have to do with the RTX 40 series?
At the launch of the RTX 4090/RTX 4080, many Nvidia showed off their new RTX Remix app. Perhaps not unjustly – if it works as intended, RTX Remix can be a very big problem for modifying older PC games.
The idea is that RTX Remix uses Ada Lovelace’s AI/machine learning features to massively streamline the process of creating HD mod packs for older games. It allows modders to import a game into the app and “capture” assets (such as textures, geometry, and lighting) in a few clicks. These assets are then automatically converted to the USD (Universal Scene Description) file format, allowing the modder to make changes using RTX Remix’s various tools.
These include AI-powered texture tools that can upscale textures up to 4x, or detect what kind of material an old texture should represent – wood, metal, etc. – and automatically give them an appropriate, more detailed roughness or shine. Ray tracing effects, Nvidia Reflex and DLSS 3 support can all be applied as well, and by synchronizing RTX Remix with other apps such as Blender, Maya, Photoshop and Adobe Substance, it is possible to create entirely new models and use them inside form to see the mod in real time.
Once completed, RTX Remix can export everything as a mod pack, which only needs to be put in the same folder as the game’s .exe to start. Plenty of games could potentially work with this: the only requirements, Nvidia says, is that they be based on DirectX 8 or 9 and use a fixed-function graphics pipeline.
Here’s Morrowind before and after the Remix treatment. Portal with RTX, seen above, was also made with Remix and is launching on Steam as free Portal DLC.
So where does the RTX 40 series come in specifically? For starters, you’ll need one of these Ada Lovelace GPUs to run RTX Remix; it is not available for older Nvidia graphics cards. And while the resulting mod packs don’t have such hardware requirements, features like ray tracing and DLSS 3 clearly encourage the use (or, let’s face it, purchase) of an RTX 4090 or RTX 4080 on the player’s side.
Will most modders want to make conversions that only owners of this brand new, crushingly expensive line of graphics cards can fully enjoy? I don’t know, although something that streamlines a tedious process like game modding is likely to attract as many eager new users as it does longtime veterans. Nvidia also talks to Nexus Mods and Mod DB about hosting Remix creations, so promoting and finding these mods shouldn’t be more difficult than handcrafted.