Why are Intel x87 registers 80 bits wide?

Question

Why use such a "weird" register size? Any documentation on why it is not preferable to use 64 or 128 bits for these registers?

Answer 1

On the page

Double precision floating point numbers are represented in 64 bits. You would like a few more bits to get higher precision for intermediate results, but it would be superfluous to use the 128-bit type when you want to get only 64 bits in the final result.

80 bits is a fairly even number of bits that is greater than 64 bits.

Consider that the data bus at the time these standards were set was 8 or 16 bits, not 32 or 64 bits, as it is today. If the standard were written today, then 96 bits would be a more reasonable number, or perhaps the data would be transmitted as 128 bits, even if all these bits would not be used in the calculations.