How to classify microprocessor

Question

Hi I am new in embedded system. I do not know the true reason we classify microprocessor into 8 bit, 16 bit, 32 bit.

In a document I read, it explained it is because of number of the bit we used to number the address of register. But I think it is not true, because if we need 32 bit to number the register address of a processor so we must have more than 2³² registers. It seem nonsense, it is too much register. So I think maybe, it is depended on the size of register or maybe the size of bus or the number of the bit, which microprocessor can work with a time.

Please help me to clarify this issued.

If you have "read" something on the subject you should include a citation, so we can see whether it is just wrong or you have misinterpreted it. — Clifford

Clifford Clifford · Accepted Answer · 2019-12-23T16:58:50

It is clear that you have either misunderstood your reference, or it is poorly worded. It should presumably state:

... number of the bits used for an address register

This means that the address range of the processor is then 2ⁿ, so perhaps your reference is referring to memory locations rather than registers.

i.e. it refers to the bit-width of a register, not the enumeration of a register.

However I would suggest that data path width is the more common and useful measure of processor architecture by "bit-width". For example 8-bit processors commonly have 16 bit address buses, and 16 bit address registers. And 16-bit 8086 devices use two 16 bit registers (32 bits) to represent a 20-bit address, but it is neither a 20 nor 32 -bit processor. 32 and 64 bit processors tend to have equal address and data register widths, which may be the cause of this erroneous statement.

As described here, the natural size of an integer (i.e. the integer size that single machine instructions take as operands) is the usual method of classification in this context.

How to classify microprocessor

2 Answers