NVIDIA recently released two new high-end graphics cards for the RTX 3080 Ti and RTX 3070 Ti, the former of which uses the GA103 core for the first time, and the latter continues to be the GA104 core.
The name GA103 appeared as early as May 2019, but it has only now landed.
It is specially designed for mobile terminals, it is still an 8nm process, Ampere architecture, 7424 stream processors, but it is not fully open, and the full blood version should have 7680.
The memory is 256-bit wide, but it is said that there are actually ten 32-bit controllers, which is physically as high as 320-bit, but it has yet to be confirmed.
Now, for the first time, we see a real photo of the GA103 core, the full number is "GN20-E8-A1".
After measurement, its area is about 496 square millimeters, which is 26.5% larger than GA104 and has 25% more cores.
It's unclear if the GA103 core will land on the desktop, and if so, will it be called the RTX 3080 SUPER?
In addition, the smallest core of the Ampere family, the GA107 used on the graphics card of the RTX 3050 Ti and RTX 3050 notebook, finally has the first photo.
However, it is not an informal retail version, but it is still a QS sample with the surface number "N20Z-GS1-A1".
At this point, the six cores of the NVIDIA Ampere family have finally come together, including the consumer GA102, GA103, GA104, GA106, GA107, all of which are 8nm processes, as well as GA100 for data centers, high-performance computing, the only 7nm process, 54.2 billion transistors, 826 square millimeters.