The cellular RTX 3080 ought to thus function a complete of 6,144 CUDA cores, as current within the fully-enabled GA-104 chip (examine that to the 5,888 CUDA cores out there on the desktop RTX 3070, and the 8,704 CUDA cores out there on the RTX 3080). These CUDA cores can be clocked at as much as 1.7 GHz. The reminiscence bus must also see a minimize all the way down to 256-bit, which might permit NVIDIA to distribute as many as four variations of the RTX 3080 cellular: Max-Q (TGP 80-90 W), Max-P (TGP 115-150 W), with both Eight GB or 16 GB of GDDR6 reminiscence. The RTX 3070 cellular retains the GA-104 chip, 256-bit bus and GDDR6 reminiscence subsystem (apparently with solely Eight GB reminiscence pool out there), however additional cuts down CUDA cores to five,120 (Max-Q TGP 80-90 W, Max-P TGP 115-150 W). Finally, the RTX 3060 cellular ought to make use of the GA106 chip, arrange with 3,072 out there CUDA cores and a 192-bit reminiscence bus throughout its 6 GB of GDDR6 VRAM pool (Max-Q TGP 60-70 W), Max-P (TGP 80-115 W). Expect these specs to be confirmed (or not) come January 12th.
