At long last, the GeForce RTX 40 series has been revealed, but in true Nvidia fashion, CEO Jensen Huang didn’t delve into raw speeds and feeds while introducing the RTX 4090 and 4080 during his GTC keynote. Heck, Huang barely talked about gaming performance. Fortunately, Nvidia’s website includes key technical details missing from the presentation, details that confirm several deeply interesting tidbits about these new “Ada Lovelace” GPU-driven graphics cards.
Here are six must-have details about the GeForce RTX 40 series that No listen during Jensen’s reveal. For full information on Nvidia’s new graphics cards, check out our coverage of the GeForce RTX 4090 and 4080 announcement.
One name, two very different RTX 4080
In a deeply anti-consumer move, Nvidia is releasing a pair of GeForce RTX 4080 offerings, which apparently differ in memory capacity. Nvidia just gave them the names “16GB RTX 4080” and “12GB RTX 4080” after all, but though Huang didn’t mention it, these two “4080”s will also offer wildly different base performance.
As you can see from the chart above, the 12GB GeForce RTX 4080 not only has less memory, it also has a much narrower 192-bit bus and a huge 21 percent less CUDA kernels. That means it will be significantly slower than the 16GB model, more often than not. Calling these two very different GPUs by the same name is sure to cause confusion among graphics card buyers.
The 12GB RTX 4080 may not have a Founders Edition model
Speaking of which, scrolling through the RTX 4080 spec sheet revealed this interesting additional tidbit. While the GeForce RTX 4080 16GB lists specific length, width, and slot measurements for Nvidia’s Founders Edition version, the GeForce RTX 4080 12GB simply says “varies by manufacturer.” It doesn’t look like the 12GB model is getting Nvidia’s custom Founders Edition treatment. (That sound you hear is the EVGA ghost screaming.)
GeForce RTX 40 series supports AV1 encoding
AV1 encoding is the holy grail for media creators, delivering much better images with much lower bandwidth needs. Intel beat out AMD and Nvidia by including AV1 encoding on their Arc graphics cards, but footnotes to Nvidia’s RTX 4090 and 4080 spec sheets reveal that Team Green will also support AV1 encoding this generation (AV1 decoding already was compatible).
“The GeForce RTX 4090 and GeForce RTX 4080 graphics cards feature two of our new 8th Generation NVIDIA Encoders (NVENC), now with support for AV1 encoding, enabling a host of new possibilities for live streamers, video editors and video callers”, the crowded RTX 40-Series Ad Booth.
You read right, not one, but two NVENC encoders support AV1 encoding on RTX 40 series graphics cards. Streamers, take note.
DLSS 3 won’t come to older GPUs
It is obvious that next generation GPUs will not include next generation hardware encoders. But Nvidia’s much-hyped DLSS 3 software feature won’t be coming to RTX 20 and 30 series cards either, despite DLSS 2.0 and 2.1 being ported to both. Nvidia’s GPU generation comparison page lists DLSS 3 on the RTX 40 series only, while previous generation RTX offerings remain listed with DLSS 2.0.
Nvidia confirmed that DLSS 3 will not be backward compatible in a statement to The Verge. “DLSS 3 works with the new 4th generation Tensor Cores and Optical Flow Accelerator on GeForce RTX 40-series GPUs” and therefore will not come to previous generations of RTX, Nvidia spokesman Benjamin said. Berraondo, to the site.
You’re gonna need a bigger boat
The GeForce RTX 4090 doesn’t draw more than 600 watts as some early rumors suggested, at least not in Nvidia’s Founders Edition iteration, but does it they come rated at 450W, with a recommended minimum PSU of 850W. That matches the flagship RTX 3090 Ti from the last generation. As we said then: you’re going to need a bigger boat. And by ship, we mean power supply. However, it’s not all expensive news. While the RTX 40 series is compatible with the 12VHPWR pin headers found on newer generation ATX 3.0 power supplies, you’ll also be able to use three or four standard 8-pin power connectors if you already have a beefy power supply.
NVLink is dead for GeForce
Nvidia’s SLI tech for multi-GPU setups has been dead for a while, but the “NVLink” tech that replaced it stuck around on last-gen RTX 3090 GPUs. no more The GeForce RTX 4090 and 4080 spec pages explicitly say that NVLink is No supported this generation. Pour one.
That’s it for these buried treasures. We’re sure to learn even more about the GeForce RTX 40 series in the coming days and weeks, but once again, for full information on Nvidia’s new graphics cards, check out our coverage of the GeForce RTX 4090 and 4080 announcement.