When the Nvidia GeForce RTX 4080 came out in November 2022 it stuck out for being nearly as expensive as the RTX 4090 but significantly slower. Fast forward to January 2024 and Nvidia is launching the RTX 4080 Super, replacing the original RTX 4080. Not only does this new graphics card feature slightly better specs, but it's actually cheaper, launching at $999 rather than the $1,199 of the original 4080.
That alone is a step in the right direction, but anyone hoping for an exciting launch that puts the original 4080 in its place is going to be disappointed. While it is technically better than the RTX 4080, this is more a price cut than a performance jump.
Finally a Better Price
The most hilarious thing about the RTX 4000 series launch is that the RTX 4090, a $1,599 graphics card, was better value than the RTX 4080, which launched at $1,199. The new RTX 4080 Super fixes that problem a bit, but it's still much more expensive than the RTX 3080, which launched in 2020 for $699. It's still a step in the right direction, though.
The high price of the original RTX 4080 made it incredibly hard to recommend, but this new price is much easier to swallow, assuming third party GPU manufacturers actually release enough cards at the new lower price. So, while I would have loved to see an Nvidia GeForce RTX 4080 Super that wiped the floor with the RTX 4080 while also being cheaper, having essentially the same graphics card in a more affordable package is alright by me, I just wish this would have been the price the RTX 4080 originally launched at.
Nvidia GeForce RTX 4080 Super – Specs and Design
Just like the rest of the RTX 4000 lineup, the RTX 4080 Super is built on the Ada Lovelace graphics microarchitecture. This architecture is based on a 4nm process and features fourth-generation Turing cores for AI acceleration and third-generation RT cores for, well, ray tracing. The RTX 4080 Super itself features 80 Streaming Multiprocessors (SMs) combining for a total of 10,240 CUDA cores, 320 Tensor cores and 80 RT cores.
The RTX 4080, in comparison, featured 9,728 CUDA cores across 76 SMs. That represents a 5% increase in silicon. As for memory, the Nvidia RTX 4080 Super sports the same 16GB of VRAM on a 256-bit bus. This means you're paying about $200 less for four more SMs, which of course means more theoretical performance at a cheaper price.
Just like the original RTX 4080, the RTX 4080 Super Founder's Edition is a triple-slot card and requires either a PCIe gen 5 cable or 3 PCIe 8-pin cables – you'll just have to use the included adapter for the latter option.
What's nice is that even though Nvidia stuffed even more hardware into the RTX 4080 Super, it doesn't require significantly more power to run. Throughout my testing the RTX 4080 Super peaked at 316W of power draw, compared to the 308W of the original RTX 4080, which basically falls into the margin of error. And, while the shroud might look a bit different, it's functionally the same, which means temperatures were nice and breezy, peaking at just 64°C.
The RTX 4080 Super also shares the same array of display outputs as any other high-end RTX 4000 series card, with three DisplayPort and one HDMI. That's pretty standard, but with how great these cards are at creative workloads, it would be awesome to see Team Green work USB-C back into these cards.
Nvidia GeForce RTX 4080 Super – Features and Software
At this point, it's impossible to talk about Nvidia's graphics cards without talking about the swath of exclusive technology Team Green's GPUs have access to.
We are thankfully at the point where every GPU has access to some kind of upscaling tech – FidelityFX Super Resolution from AMD and XeSS from Intel – but neither of those are quite as good as Nvidia's DLSS. This was true when DLSS first launched, but the technology has only improved in the six years since its debut with the RTX 2080.
DLSS 3.0, the new generation deep learning super-sampling tech for the RTX 4000s, doesn't only upscale your game from a lower resolution back up to your native resolution. It also has Frame Generation, which uses the AI-accelerating Tensor Cores to constantly read image and motion data from your game in order to generate new frames. This completely replaces the traditional rendering pipeline where your GPU is constantly waiting for image information from your CPU, and as such drastically improves your frame rate.
It's not perfect, and really works best if you're already getting a decent framerate in your game, which helps the algorithm accurately predict motion in the game world, but it's a nice feature that no other brand of graphics card has access to right now.
Honestly, it would be best if every major GPU manufacturer could use the neural processing hardware that exists in every mainstream graphics card right now to create an upscaling technology like DLSS. But until that happens, Nvidia simply offers the best upscaling tech in GPUs right now.
Nvidia Geforce RTX 4080 Super – Performance
The Nvidia GeForce RTX 4080 Super, just like the original RTX 4080, is a 4K graphics card through and through. In every test, this GPU excels at that high resolution with plenty of room to spare, especially if you utilize Nvidia's DLSS.
For instance, in Cyberpunk 2077 with the Ray Tracing Ultra preset and DLSS set to "performance", the RTX 4080 Super hits 84 fps at 4K. Turn on the new Frame Generation tech and that framerate jumps up to 111 fps.
However, while performance is excellent, the RTX 4080 achieves basically the same results, scoring 80 fps at 4K without Frame Generation and the same 111 fps with it enabled. So, at best, you get a 5% increase in performance – but that's not bad when you consider this is a cheaper card.
Even in games without DLSS or ray tracing, like Total War: Warhammer 3, the RTX 4080 Super is a monster at 4K, scoring 79 fps, compared to 76 from the RTX 4080 and 86 from the AMD Radeon RX 7900 XTX.
In synthetic benchmarks the RTX 4080 Super is able to stretch its legs a bit. In the Port Royal benchmark, which tests ray tracing performance, the RTX 4080 Super scores a whopping 18,580 points, compared to 17,819 from the RTX 4080. That's a 4% lead. And in Speed Way, which tests DirectX 12 Ultimate performance, the RTX 4080 Super manages 7,493 points, compared to 7,314 from the RTX 4080 and 6,222 from the Radeon RX 7900 XTX.
The elephant in the room, however, is the gap between the RTX 4080 Super and the RTX 4090. When the RTX 4080 first came out, it struggled to perform close enough to the RTX 4090 to justify its $1200 price tag. And while the performance hasn't improved by that much, lowering the price makes the RTX 4080 Super a much more appealing purchase. The RTX 4090 is still about 20-25% faster in pretty much every test I ran, but instead of just being 30% more expensive than the RTX 4080, there's now a much steeper 60% price difference between the 4090 and the 4080 Super.
This means the meme of the RTX 4090 being a better value than the RTX 4080 is officially over. Yeah, you're going to be getting significantly better performance if you spend almost $2000 on an RTX 4090, but you're getting better performance per dollar spent from the RTX 4080 Super. Finally.
