This is related to the word length of the computer.
For example, 8086 is 16 bit CPU. If an integer is placed in an 8-bit register such as AL (the lower 8 bits of register AX can be regarded as an independent register), then the range of data it can handle is-128 to 127 (that is, 2 7). The most significant bit is the sign bit, so only 7 bits can be used. Similarly:
If it is placed in the 16-bit register AX, the data range it can handle is -32768~32767 (2 15).
If it is a 32-bit CPU, its data register AX is 32 bits, namely 4 bytes.
If a data type is defined as 4 bytes, it can be filled with AX, and the processing range is-231~ 231-1.
As for why the scope is asymmetric. It's about supplements.
Of course, a 16-bit CPU can also process 32-bit data, so it is necessary to put the high 16 bits of a number in AX and the low 16 bits in DX, and the problem of carrying is more complicated.
I wonder if my personal experience can help you.