Either using CuBitcrack or ClBitcrack for;

1x RTX™ A6000 (total 48 GB GPU Memory, 10,752 CUDA cores)
2x RTX™ 3090 (total 48 GB GPU Memory, 10,496 CUDA cores)

Same total GPU memory, but which one GPU can get more MKey/s?

Any advice speed in MKey/s, base on GPU Memory or CUDA cores?