A new rumor from NVIDIA concerns the card’s entry-level GPUs. In fact, the company may well be exploiting an Ada Lovelace chip in an older-generation entry-level card. It would appear that an RTX 3050 A in AD106 is in the pipeline!
RTX 3050 A: an older-generation card with a current GPU!
Broadly speaking, the card uses a current GPU, an AD106 to be precise. This is a chip already found in the RTX 4060 Ti and RTX 4070, as well as the mobile version of the 4070.
Compared with the RTX 3050 using a GA107 GPU, this GPU offers greater transistor density and more recent functionalities… Not to mention a potentially larger configuration. Nevertheless, don’t expect an overpowered 3050, no! NVIDIA’s aim is probably to recycle faulty GPUs by integrating them into an entry-level board. The configuration would be tailored to match the performance of a 3050, with a small gain of no more than 5%.
As for functionalities, we’ve learned that the chameleon can disable what’s necessary to take advantage of DLSS 3 – no kidding!
However, the benefits of TSMC’s 4N etching will be felt on power consumption, which will be reduced, as will heating. For an entry-level laptop chip, this would be a good thing.