Nvidia’s rumored RTX Super graphics card refreshes could maintain the same power usage as the existing models that they will be supercharged versions of – with implications for performance therein.
Of course, this is just speculation aired on X (formerly Twitter) from one of the usual suspects, leaker Kopite7kimi, so take it with a great deal of caution.
They will have the same power consumption than non-super models. https://t.co/1vG6lsO4KxNovember 2, 2023
As you can see in the tweet, the leaker states that the Super refreshes will have the same power usage as their counterparts, but for one exception noted in a further tweet.
That’s the lowest-end refresh, the RTX 4070 Super, which will supposedly have a 220W power consumption rather than the 200W seen with the vanilla RTX 4070.
Apparently the purported RTX 4070 Ti Super and RTX 4080 Super will run with the same (maximum) power drain as the existing RTX 4070 Ti and RTX 4080, meaning 285W and 320W respectively.
Analysis: Weak sauce worries about the RTX 4080 Super
Okay, so if this turns out to be correct – and we’ll come back to discuss that point in a moment – what does it mean? Namely that Nvidia might be pushing clocks a bit harder with the RTX 4070 Super (hence more wattage), but otherwise, there are no plans to really drive forward with meaningfully higher clocks for the other Super variants.
Is that disappointing? Well, we’ve certainly heard rumblings from the rumor mill in the past that the inbound Super graphics cards will benefit from faster clock speeds, so in that respect, it is.
However, clocks obviously aren’t the only important aspect of performance – they’re just one piece of a larger puzzle. We’re looking at other upgrades for these refreshes, such as possible process optimizations, beefier VRAM perhaps in some cases, and moreover increases in CUDA core counts which are definitely coming.
Some core count boosts are going to be substantial as per previous rumors from Kopite7kimi, notably the RTX 4070 Super, which will get a 22% increase in cores in theory compared to the RTX 4070 (and faster clocks, as this is the card to get something of a power boost). The RTX 4070 Ti Super is also looking at a solid 10% gain in cores (over the existing Ti), if the leaker is on the money.
Where the potential disappointment lies is with the RTX 4080 Super, which if using AD103 (the idea of repurposing the higher-tier AD102 chip now seemingly abandoned by the rumor mill, for various reasons that make plenty of sense) can only go to the full loadout of that GPU – a 5% CUDA core boost.
That’s not much, and as AD103 also rules out VRAM boosts (due to its spec – for the previously rumored 20GB of video RAM, AD102 would be required), plus we apparently don’t have much movement on clock speeds either, this leaves the RTX 4080 Super looking like a potentially lukewarm upgrade from the base RTX 4080.
And that seems odd, really, as part of Nvidia’s strategy with these refreshes is, as far as we can tell, to address the RTX 4080’s widely reported poor (or even abysmal) sales. So, why make the RTX 4080 Super the weakest variant?
Well, maybe because it’s outright replacing the RTX 4080, as per rumors that the existing card might be scrapped, and the 4080 Super could arrive with extremely attractive pricing, if another rumor is right. So it’d make a good value proposition, still, in that case.
Overall, though, this latest buzz from the grapevine is a little worrying for the top-end Super refresh. However, coming back to the accuracy of this latest spillage, we must bear in mind with all these leaks that Nvidia hasn’t yet come close to making the final decisions on power consumption and clock speeds, or even core loadouts and model names.
There’s still room to maneuver before a launch in early 2024, and how Nvidia pitches these refreshed graphics cards will depend on how the market shifts over the next few months. And indeed what AMD is rumored to be preparing as a response to these Super variants (which will doubtless be leaked at some stage).