NVIDIA will launch the RTX 3060 by the end of February. It will be based on the GA106 GPU, which is the fourth GPU from the NVIDIA Ampere family.
Below are the pictures for the GPU codenamed GA106-400, which is the actual full chip. Initially, NVIDIA had plans to launch the RTX 3060 operating on GA106-400 with 3840 CUDA cores, but as the date came closer, NVIDIA opted for the GA106-300 GPU with 3584 cores; hence the picture is named as GA106-400. The picture shows an engineering sample.
In other words, this board could easily be adapted for RTX 3060 Ti or even RTX 3050 Ti, which is also rumored to use GA106 GPU in the future.
The RTX 3060 will operate on K4ZAF3258M-HC16 memory clocked at 16 Gbps. The memory will have 6 modules where each module will occupy 2 gigs of RAM. The GA104 and GA106 share the same memory; hence they will be pin-compatible, reducing the development cost the AIB will incur, and two unoccupied slots are visible in the picture. The RTX 3060 Ti based on the GA104 GPU will occupy all eight slots, but each slot will have 1 GB RAM and 14Gbps operating speed.
The GPU has to die size of 276 mm2, which appears to be quite small when compared to the Turing GPUs; the GA106 has a significant size difference than TU106 (445 mm2) and a negligible size difference from TU116 (284 mm2).
Via Videocardz