As you may know, both AMD and NVIDIA have projects in the works based on MCM (Multi-Chip Module) designs. In fact, it is very likely that both manufacturers will start releasing graphic products based on the same design philosophy as early as 2022. However, despite the huge impact of this change, only AMD will be adopting this type of technology for the PC gaming world as of now.
Yes, NVIDIA “already” relies on MCM, but only for large data centers and supercomputers.
However, in case you don’t understand what MCM is, this is a design philosophy very similar to that of AMD already in their Ryzen processors. In other words, we will have processing units in separate modules within the same “package”. A solution that is effectively used to significantly increase processing power without unduly increasing production complexity.
NVIDIA will be using an MCM design as early as 2022. But not for the PC
As a result, NVIDIA has just announced its GPU Tech 2022 conference, now in physical form, where NVIDIA Founder and CEO (Jensen Huang) will present the new graphics architectures. Especially one in which high hopes are raised, the ‘Hopper’! Of course, it is the company’s first graphics architecture to feature a multi-chip module design.
However, that announcement also seems to make it clear that the architecture will be more for the HPC and AI world, with the Ada Lovelace monolithic architecture handling the gaming PC (RTX 40) market.
Before we start discussing what might or might not happen, we already know that we will be hosting a hybrid event here in March that will begin with a keynote with Jensen Huang. However, March is already a time when the gaming world is starting to indulge and of course a new generation of graphics cards is being considered, so it would not be surprising to see some “clues” about the RTX 40.