Rumours are circulating about a potential beefed-up version of Nvidia’s graphics cards in the form of the GeForce RTX 4080 Ti, slated to make its debut in the early months of 2024. While the official nomenclature remains undecided, the GPU is said to employ a trimmed-down version of the powerful AD102 GPU, which currently powers the RTX 4090. With a purported power limit of under 450W and a price range similar to that of the RTX 4080, the prospect of this new release has stirred considerable interest.
While the source of this information is attributed to social media chatter, it is essential to approach it with caution, given the absence of concrete evidence. Nonetheless, it’s worth dissecting the claims to gain a better understanding of what might be on the horizon.
Firstly, let’s consider the use of the AD102 GPU. The existing GeForce RTX 4080 utilizes its smaller counterpart, the AD103, equipped with a maximum of 80 SMs (Streaming Multiprocessors) and a 256-bit memory bus. However, the current RTX 4080 only disables four SMs, leaving the rest of the die fully operational. Consequently, there’s limited potential for creating a significantly faster product using this chip, as a fully enabled AD103 operating at the same clock speed as the one in the RTX 4080 would yield only marginal performance gains.
On the other hand, the AD102 boasts more substantial capabilities, featuring up to 144 SMs and a 384-bit memory bus. The RTX 4090 employs a version with 128 SMs enabled. Nvidia could potentially utilize an AD102 variant with, for instance, 100 SMs for the RTX 4080 Ti. This particular chip is already in use in the RTX 5000 Ada Generation workstation card, operating at a similar clock speed to that of the RTX 4080 but with a significantly lower power limit.
Such a configuration could potentially result in the Ti version being up to 25% faster than the original model, especially if it employs a 320-bit memory bus configuration and incorporates 20GB of VRAM. Nevertheless, it’s crucial to note that the RTX 4080 hasn’t enjoyed the same level of success as its predecessors due to its significantly higher price. While an updated Ada Lovelace card may not be more affordable, if it maintains a similar price range to the original or is only slightly more expensive, a Ti variant could potentially fare better.
However, it remains to be seen whether the market will exhibit significant interest in the RTX 4080 Ti. Given the substantial price difference between the RTX 4080 and the RTX 4090, it is likely that consumers who can afford the former might prefer to invest a bit more for the latter. This pricing dilemma poses a challenge for Nvidia.
Additionally, the issue extends beyond just the RTX 4080 Ti. Nvidia must navigate the complexities of its entire RTX 40-series lineup. The existence of RTX 4060/4070 Ti models, it raises questions about what other models are left to be released. The “Super” moniker, previously used sparingly in the Turing series, could potentially lead to confusion among users who currently own Turing-based Super models and are looking to upgrade.
Nvidia faces the challenge of managing its inventory of Ada Lovelace GPUs. These GPUs go through a process called chip binning, where those that do not meet the requirements for a current model are set aside for future use. In the past, Nvidia employed them in Ti or Super versions or used high-end chips in lower-end models with various parts disabled, all without changing the product name.
However, with the existence of RTX 4060/4070 Ti models, the question arises: if Nvidia releases an RTX 4080 Ti in early 2024, what’s left to launch? Perhaps an RTX 4050 is in the cards, but this is purely speculative. Nvidia finds itself in a situation where it must address the complexities of product names and specifications in the Ada Lovelace generation of GeForce cards.
As next year approaches, Nvidia’s options seem limited to refreshed versions or maintaining the status quo, as its next generation of GPUs is not expected to arrive until well into 2025. While these challenges lie squarely in Nvidia’s domain, it underscore the importance of adopting a consistent GPU tiering system to avoid confusion among consumers. The Ada Lovelace generation has, in this regard, been a perplexing journey for Nvidia.
