Quoted from http://jheriko-rtw.blogspot.com/2009/04/why-do-we-have-f2xm1.html?showComment=1249084660633#c6336764748095052620
F2XM1 is a floating point assembler instruction for Intel CPUs. It is one of several which seem to be there to allow the calculation of several of the common "higher" functions, like pow(x, y) or log(x). However, I am always confused why it subtracts one after finding the power of 2 and why its limited from -1 to 1, since it doesn't seem to help anything much at all...
If anyone has any comments or suggestions about what use F2XM1 might have, and why Intel's FPU architecture also lacks an instruction to perform 2^x without subtracting anything (I know of at least one other FPU instruction set that does, and has no input range limitations).
Link to Original Source