This video card model is based on NVIDIA’s GA104 silicon, featuring the Ampere architecture and produced by Samsung’s 8 nm process. In this 392 mm2 die we have 4864 CUDA Cores, 152 Tensor Cores (3rd Gen), 38 RT Cores (Ray Tracing Cores) (2nd Gen), 152 TMU’s and 80 ROP’s. It has 8GB of GDDR6 memory at 14GHz on a 256-bit memory bus, which provide 448GB/s of total memory bandwidth. The TGP for this part is 200W. As we mentioned previously our video card has a roughly 3% boost clock increase over the 1665 MHz boost clock of the NVIDIA GeForce RTX 3060 Ti Founders Edition, which equates to a factory boost clock of 1710 MHz. You also gain access to features like NVIDIA REFLEX, GeForce Broadcast, EVGA Precision X1, as well as GeForce RTX IO.



Link: The FPS Review