How fast does Nvidia’s new upper-class model RTX 4080 Super work? How much power does it need and how loud does it work? The test gives answers!
Table of contents
- RTX 4080 Super in the test: 10,240 computing units
- TDP increased by 20 watts
- Moderate speed advantage
- Power consumption almost identical
- RTX 4080 Super in the test: conclusion
With the RTX 4080 Super, Nvidia is now the third Super chip from the RTX 4000 series the Graphics cardsavailable to manufacturers. And in the tests it came dangerously close to the top graphics processor RTX 4090, as the test of the RTX 4080 Super shows.
RTX 4080 Super in the test: 10,240 computing units
At the top end, the one after that also receives RTX 4090 second strongest graphics chip RTX 4080 from the Ada Lovelace series, a fresh cell treatment. The 4080 Super can now distribute calculations to 10,240 Cuda units instead of the previous 9,728. Chip type (AD103) and size (379 mm2) are identical, but Nvidia has given the 4080 Super model an additional 16 tensor cores and four extra ray tracing cores for calculating ray tracing effects such as reflections and fine shadows.
TDP increased by 20 watts
In addition, the number of texture units responsible for calculating the surfaces of characters, objects and environments in games increases from 304 to 320. The consumption limit (TDP) increases Nvidia from 320 to 340 watts. Important: The manufacturer will stop producing the model without Super in the name; in the future, only RTX 4080 Super variants will be available in stores.
Nvidia GeForce RTX 4000 series: technical data
Nvidia GeForce RTX 4000 series: technical data
GeForce RTX 4090 | GeForce RTX 4080 Super | GeForce RTX 4080 | GeForce RTX 4070 Ti Super | GeForce RTX 4070 Ti | GeForce RTX 4070 Super | GeForce RTX 4070 | GeForce RTX 4060 Ti 16GB | GeForce RTX 4060 Ti 8GB | GeForce RTX 4060 | |
---|---|---|---|---|---|---|---|---|---|---|
production
|
5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC | 5 nanometers at TSMC |
GPU type |
AD102 | AD103 | AD103 | AD103 | AD104 | AD104 | AD104 | AD106 | AD106 | AD107 |
Chip size |
608mm² | 379mm² | 379mm² | 379mm² | 294mm² | 294mm² | 294mm² | 190mm² | 190mm² | 146mm² |
Transistors |
76.3 billion | 45.9 billion | 45.9 billion | 45.9 billion | 35.8 billion | 35.8 billion | 35.8 billion | 22.9 billion | 22.9 billion | 18.9 billion |
Shader models |
128 | 80 | 76 | 66 | 60 | 56 | 46 | 34 | 34 | 24 |
Cuda cores |
16,384 | 10,240 | 9,728 | 8,448 | 7,680 | 7,168 | 5,888 | 4,352 | 4,352 | 3,072 |
Base clock |
2,230MHz | 2,205MHz | 2,205MHz | 2,340MHz | 2,310MHz | 1,980MHz | 1,920MHz | 2,310MHz | 2,310MHz | 1,830MHz |
Boost clock |
2,520MHz | 2,505MHz | 2,505MHz | 2,610MHz | 2,610MHz | 2,475MHz | 2,475MHz | 2,535MHz | 2,535MHz | 2,460MHz |
Tensor cores |
512 (4th gen.) | 320 (4th gen.) | 304 (4th gen.) | 264 (4th gen.) | 240 (4th gen.) | 224 (4th gen.) | 184 (4th gen.) | 128 (4th gen.) | 128 (4th gen.) | 96 (4th gen.) |
Ray tracing cores |
128 (3rd gen.) | 80 (3rd gen.) | 76 (3rd gen.) | 66 (3rd gen.) | 60 (3rd gen.) | 56 (3rd gen.) | 46 (3rd gen.) | 32 (3rd gen.) | 32 (3rd gen.) | 24 (3rd gen.) |
Texture units |
512 | 320 | 304 | 264 | 240 | 224 | 184 | 128 | 128 | 128 |
Graphics memory |
24GB GDDR6X | 16GB GDDR6X | 16GB GDDR6X | 16GB GDDR6X | 12GB GDDR6X | 12GB GDDR6X | 12GB GDDR6X | 16GB GDDR6 | 8GB GDDR6 | 8GB GDDR6 |
Storage interface |
384 bits | 256 bits | 256 bits | 256 bits | 192 bits | 192 bits | 192 bits | 128 bits | 128 bits | 128 bits |
TDP |
450 watts | 340 watts | 320 watts | 320 watts | 285 watts | 225 watts | 200 watts | 165 watts | 160 watts | 115 watts |
Moderate speed advantage
Given the serious technical innovations, the RTX 4080 Super should offer a major speed advantage compared to the normal RTX 4080. But it doesn’t, as this test shows. Because Nvidia appears to have balanced everything so finely that the RTX 4080 Super doesn’t follow too closely on the heels of the top model RTX 4090. The result: The increase in performance in direct comparison to the normal RTX 4080 is moderate – on average, the super model was five percent faster in the speed tests (see speed comparison below).
Power consumption almost identical
Good: Despite the increased performance, the new RTX 4080 Super required just as much energy as the standard model. At its peak, it drew 308 watts from the socket, but for Internet and office applications and when playing 4K videos it was only 17 watts. Also good: While the fans were clearly audible, especially under full load (5.4 sones), the tested Zotac GeForce RTX 4080 Super Trinity Black worked silently when watching films and working with office applications.
RTX 4080 Super in the test: conclusion
At the time of testing, the new Super version of the RTX 4080 was around 150 euros more expensive than the standard model without Super in the name: 1,350 euros for the tested Zotac GeForce RTX 4080 Super Trinity Black instead of around 1,200 euros. That’s a lot of wood for a speed advantage of just five percent. Given the key technical data, the 4080 Super could definitely work a bit faster. But that’s probably not possible, because the RTX 4080 Super has to maintain the necessary respect distance from the much more expensive RTX 4090. Nevertheless: The RTX 4080 Super completed the speed tests in a hurry and only used marginally more energy. But if you already have an RTX 4080 graphics card in your computer, you don’t need the new super model. However, if you are thinking about buying a 4080 graphics card, you should go for the Super version. Because Nvidia will refine the driver over the next few years and thus extract even more performance. This is unlikely with the discontinued model RTX 4080 without a super abbreviation.