While the GeForce RTX 40XX launch came with many impressive performance figures earlier this week, it didn’t escape people’s attention that, in terms of technical specifications, Nvidia was notably light on the details. This did, of course, lead to the suggestion that Nvidia might have been getting a little coy with the community. You know, blinding us with DLSS3/ray-tracing razzle-dazzle gaming performance while poking the actual technical specifications of the graphics cards to a somewhat darker less obvious corner.
Following a report via TechPowerUp, however, the complete design specifications of all three GPUs have now been confirmed, and, if nothing else, all it might end up suggesting is that the 40XX series might be one of the busiest and confusing Nvidia has ever created.
The figures above clearly provide the community with a far better picture of where the graphics cards currently confirm sit in the grand scheme of things. And more so, where future releases may look to go. Starting with the AD102 chipset though, their top spec cards are clearly more than a little impressive. With the AD102 chipset allowing up to 18,432 CUDA cores when fully unlocked, this clearly leaves some wiggle room for a future 4090 Ti given that the standard 4090 only(?) has 16,384.
When you look at the 4080, however, you really start to notice how the figures, in many respects, simply don’t make any sense when compared to the product name/skew Nvidia has chosen to use.
One of the initial reactions to the unveiling of the 4080 was that it didn’t make much sense that the 16GB version came with a $300 higher MSRP than the 12GB variant for what was essentially, or at least from what Nvidia showed at the time, a GPU that only apparently offered 4GB more of VRAM as the key difference. – With these specifications, however, it clearly highlights that while the GPUs might share the same name, they categorically do not share the same specifications. Hell, they don’t even share the same chipset! – Put simply, the 4080 16GB is a more potent GPU in practically every measurable way. – The 4080 12GB doesn’t just have less VRAM, it has notably fewer CUDA cores, fewer Tensor and RT cores, and a smaller memory bus speed.
Why is this the case though? Well, in something that seems to back up a lot of rumours which started around April earlier this year, it’s almost certainly because the 4080 12GB isn’t a 4080 at all. It’s a 4070 which Nvidia has decided to kick up the pecking order. Hell, some even think it’s a 4060! Why would they do this though? Well, one theory suggests that it’s because by calling it a 4080 they can slap an $899 MSRP on it instead of something like $699 which most would expect/anticipate for the 4070 or $499 for the 4060.
Overall though, this information shows that there are literally masses of options Nvidia may undertake with its top-spec 40XX graphics cards. From the 4080 12GB up, there’s more than enough room for the Ti models based on the chipset specifications, and who knows, perhaps even the ‘SUPER’ branding might make a return. – With Nvidia making the distinction between the 4080 16GB and 12GB so unclear, however, from a general consumer standpoint, this is surely only going to get messier and messier as further GPUs arrive!
It is, however, pretty hard to counter the argument that the 4080 12GB has been kicked up in its tier so its price tag can follow suit!
What do you think? – Let us know in the comments!
AMD has released a significant update for gamers playing Warhammer 40,000: Space Marine 2, introducing…
The AMD Ryzen 5 7600X3D is making waves in the gaming world, beating some of…
PlayStation’s newest release, Astro Bot, has become a massive hit in Japan, where it has…
A major announcement may be on the horizon for Xbox gamers. According to Jez Corden…
Larian Studios has unleashed a new wave of creativity for Baldur's Gate 3 players with…
Go beyond performance with the latest 14th Generation Intel Core processors, based on the Raptor-Lake…