NVIDIA RTX 3050 A, new SKU with AD106 ADA GPU

The company has confirmed to Notebookcheck that such GPU is indeed happening.

A quick update to our story about the supposed Ada Lovelace-based GeForce RTX 30 model: It turns out NVIDIA is indeed planning to put an Ada GPU into its old series. However, note that while the company confirms the initial specs, it has not yet confirmed it is based on the Ada architecture.

A quick recap: The RTX 3050 A (as it is named) is not a desktop GPU but a mobile variant. It clearly follows the release of the RTX 2050, which featured an Ampere GPU, not Turing like the rest of the RTX 20 series.

The new release, the RTX 3050 A (Laptop GPU), features 1792 CUDA cores, as confirmed by NVIDIA through NotebookCheck. The card is also said to feature 4GB of memory across a 64-bit memory bus and a configurable TGP between 35 and 50 watts.

Update 2024/07/26 12:08 CET: Nvidia has confirmed the GeForce RTX 3050 (titled GeForce RTX 3050 A) in question exists and will feature the following specs: 4 GB VRAM on a 64-bit bus, 35-50 Watt TDP, and 1,792 CUDA cores.

— NotebookCheck

Compared to the existing 3050 SKUs, the new model has fewer cores and a shorter memory bus. However, without data on the memory speed, bandwidth may not be significantly lower if NVIDIA used faster memory.

In comparison to the RTX 2050, the specs are similar in terms of memory. However, the ‘predecessor’ still has more memory. Given the architectural upgrade from Ampere to Ada, the new 3050 A may perform similarly, depending on the workload. This, of course, will require independent testing, but it seems that this launch has a limited scope. NVIDIA did not even make a public announcement. In fact, none of the laptop makers have announced this GPU option yet, despite the driver release over 10 days ago.

Source: NotebookCheck